Institute of Informatics>>> Department of Image Processing and Computer Graphics>>>

Magyarul

Selected Publications of the Department of Image Processing and Computer Graphics

1. Jayaram K. Udupa, Nong Sang, Laszlo G. Nyul, and Hengqing Tong, editors. MIPPR 2009: Multispectral Image Acquisition and Processing, volume 7494 of SPIE Proceedings. SPIE, Yichang, China, November 2009. [WWW] [doi:10.1117/12.839775] [bibtex-key = spie:7494]


2. Gabor T. Herman and Attila Kuba, editors. Advances in Discrete Tomography and Its Applications, Applied and Numerical Harmonic Analysis. Birkhauser, pubadd-Birk, 2007. [WWW]
   Abstract: Advances in Discrete Tomography and Its Applications is a unified presentation of new methods, algorithms, and select applications that are the foundations of multidimensional image reconstruction by discrete tomographic methods. The self-contained chapters, written by leading mathematicians, engineers, and computer scientists, present cutting-edge research and results in the field. Three main areas are covered: foundations, algorithms, and practical applications. Following an introduction that reports the recent literature of the field, the book explores various mathematical and computational problems of discrete tomography including new applications. Topics and Features: * introduction to discrete point X-rays * uniqueness and additivity in discrete tomography * network flow algorithms for discrete tomography * convex programming and variational methods * applications to electron microscopy, materials science, nondestructive testing, and diagnostic medicine Professionals, researchers, practitioners, and students in mathematics, computer imaging, biomedical imaging, computer science, and image processing will find the book to be a useful guide and reference to state-of-the-art research, methods, and applications. [bibtex-key = Herman2007]


3. Attila Kuba, Laszlo G. Nyul, and Kalman Palagyi, editors. Discrete Geometry for Computer Imagery, volume 4245 of Lecture Notes in Computer Science. Springer Verlag, October 2006. [WWW] [doi:10.1007/11907350]
   Abstract: This book constitutes the refereed proceedings of the 13th International Conference on Discrete Geometry for Computer Imagery, DGCI 2006, held in Szeged, Hungary in October 2006. The 28 revised full papers and 27 revised poster papers presented together with 2 invited papers were carefully reviewed and selected from 99 submissions. The papers are organized in topical sections on discrete geometry, discrete tomography, discrete topology, distance, image analysis, shape representation, segmentation, skeletonization, as well as surfaces and volumes. [bibtex-key = Kuba2006]


4. Vito Di Gesu and Attila Kuba, editors. IWCIA 2003-Ninth International Workshop on Combinatorial Image Analysis, volume 151 of Discrete Applied Mathematics. Elsevier, October 2005. [WWW] [bibtex-key = Gesu2005]


5. Gabor T. Herman and Attila Kuba, editors. Proceedings of the Workshop on Discrete Tomography and its Applications, volume 20 of Electronic Notes in Discrete Mathematics. Elsevier, July 2005. [WWW] [bibtex-key = Herman2005]


6. Alberto Del Lungo, Vito Di Gesu, and Attila Kuba, editors. 9th International Workshop on Combinatorial Image Analysis, volume 12. Elsevier, Palermo, Italy, March 2003. [WWW] [doi:10.1016/S1571-0653(04)00469-X]
   Abstract: The 9th International Workshop on Combinatorial Image Analysis was held between 14 and 16 May, 2003 in Palermo (Italy). Prior meetings took place in Paris (France, 1991), Ube (Japan, 1992), Washington DC (USA, 1994), Lyon (France, 1995), Hiroshima (Japan, 1997), Madras (India, 1999), Caen (France, 2000) and Philadelphia (USA, 2001). The workshop focused on digital topology, discrete tomography, array, graph and parallel grammars, cellular automata, tilings and patterns, combinatorial methods in image analysis, combinatorial problems in the discrete plane, integer-only algorithms for image analysis, and non-conventional computing for image analysis, like genetic algorithms and global optimization. We are very pleased that P. Grirtzmann (Technische Universität Munchen), Gabor T. Herman (City University of New York) and Reinhard Klette (The University of Auckland) accepted our invitations to speak at the Workshop. These proceedings contain thirty-one papers and they provide an exciting cross-section of current research in our field. The editors would like to thank the valuable work to the members of the program committee: Alfred M. Bruckstein, Technion Israel Institute of Technology, Israel Jean-Marc Chassery, TIMC-IMAG, France Katsushi Inoue, Yamaguchi University, Japan Yung Kong, City University of New York, USA Kenichi Morita, Hiroshima University, Japan Maurice Nivat, LIAFA, FranceMike Reed, Oxford University, UK Rani Siromoney, Madras Christian College, India Patrick S.P. Wang, Northeastern University, USA Gerhard Woeginger, University of Twente, The Netherlands The editors thank also for the technical help given by Cesare Valenti, University of Palermo, Italy Andrea Frosini, University of Siena, Italy Laszlo Nyul, University of Szeged, Hungary Finally, we would like to thank Peter Hammer for inviting us to publish the Workshop's proceedings online as a volume of Electronic Notes in Discrete Mathematics, and for agreeing to devote a special issue of Discrete Applied Mathematics to IWCIA'2003. The DAM special issue will contain a refereed selection of papers submitted by Workshop presenters after the meeting. [bibtex-key = Lungo2003]


7. Gabor T. Herman and Attila Kuba, editors. Discrete Tomography: Foundations, Algorithms, and Applications, Applied and Numerical Harmonic Analysis. Birkhauser, Boston, December 1999. [WWW] [bibtex-key = Herman1999]


8. Attila Kuba, Martin Samal, and Andrew Todd-Pokropek, editors. Information Processing in Medical Imaging: 16th International Conference, IPMI'99, Visegrad, Hungary, June/July 1999. Proceedings, volume 1613 of Lecture Notes in Computer Science. Springer Verlag, 1999. [WWW]
   Abstract: This book constitutes the refereed proceedings of the 16th International Conference on Information Processing in Medical Imaging, IPMI'99, held in Visegrad, Hungary in June/July 1999. The 24 revised full papers and the 28 posters presented have been carefully reviewed and selected from a total of 82 submissions. The volume addresses the full range of current topics in the area in particular new imaging techniques, 3D ultrasound and PET, segmentation, image analysis of the brain cortex, registration, feature, detection and modelling, cardiovascular image analysis, shape modelling and analysis, segmentation and detection, measurement and quantitative analysis, and analysis of image sequences and functional imaging. [bibtex-key = Kuba1999]


9. Gabor T. Herman and Attila Kuba, editors. Discrete Tomography (Special Issue), volume 9. John Wiley & Sons, Inc., 1998. [doi:10.1002/(SICI)1098-1098(1998)9:2/3<67::AID-IMA1>3.0.CO;2-L]
   Abstract: no abstract [bibtex-key = Herman1998]


10. Peter Kardos. Sufficient conditions for order-independency in sequential thinning. Acta Cybernetica, 20:87-100, 2011. [PDF]
    Abstract: The main issue of this paper is to introduce some conditions for template-based sequential thinning that are capable of producing the same skeleton for a given binary image, independent of the visiting order of object points. As an example, we introduce two order-independent thinning algorithms for 2D binary images that satisfy these conditions. [bibtex-key = K_ACTA2011]


11. Gabor Nemeth, Peter Kardos, and Kalman Palagyi. 2D parallel thinning and shrinking based on sufficient conditions for topology preservation. Acta Cybernetica, 20:125-144, 2011. [PDF]
    Abstract: Thinning and shrinking algorithms, respectively, are capable of extracting medial lines and topological kernels from digital binary objects in a topology preserving way. These topological algorithms are composed of reduction operations: object points that satisfy some topological and geometrical constraints are removed until stability is reached. In this work we present some new sufficient conditions for topology preserving parallel reductions and fifty-four new 2D parallel thinning and shrinking algorithms that are based on our conditions. The proposed thinning algorithms use five characterizations of endpoints. [bibtex-key = NKP_ACTA2011]


12. Gabor Nemeth, Peter Kardos, and Kalman Palagyi. Thinning combined with iteration-by-iteration smoothing for 3D binary images. Graphical Models, 73:335-345, November 2011. [PDF]
    Abstract: In this work we present a new thinning scheme for reducing the noise sensitivity of 3D thinning algorithms. It uses iteration-by-iteration smoothing that removes some border points that are considered as extremities. The proposed smoothing algorithm is composed of two parallel topology preserving reduction operators. An efficient implementation of our algorithm is sketched and its topological correctness for (26,6) pictures is proved. [bibtex-key = NemethKardosPalagyi2011GMOD]


13. Gabor Nemeth and Kalman Palagyi. Topology preserving parallel thinning algorithms. International Journal of Imaging Systems and Technology, 21:37-44, 2011. [PDF] [doi:DOI 10.1002/ima.20272]
    Abstract: Thinning is an iterative object reduction technique for extracting medial curves from binary objects. During a thinning process, some border points that satisfy certain topological and geometric constraints are deleted in iteration steps. Parallel thinning algorithms are composed of parallel reduction operators that delete a set of object points simultaneously. This article presents 21 parallel thinning algorithms for (8,4) binary pictures that are derived from the sufficient conditions for topology preservation accommodated to the three parallel thinning approaches. [bibtex-key = NP_IJIST2011]


14. Laszlo Varga, Peter Balazs, and Antal Nagy. Direction-dependency of binary tomographic reconstruction algorithms. Graphical Models, 73:365-375, November 2011. [PDF]
    Abstract: In this work we study the relation between the quality of a binary tomographic reconstruction and the choice of angles of the projections. We conduct experiments on a set of software phantoms by reconstructing them from different projection sets using three different discrete tomography reconstruction algorithms, and compare the accuracy of the corresponding reconstructions with suitable approaches. To validate our results for possible real-world applications, we conduct the experiments by adding random noise of different characteristics to the simulated projection data, and by applying small topological changes on the phantom images as well. In addition, we also discuss some consequences of the angle-selection dependency and possible practical applications arising from the field of non-destructive testing, too. [bibtex-key = VargaBalazsNagy2011GM]


15. Laszlo Varga, Peter Balazs, and Antal Nagy. Projection selection dependency in binary tomography. Acta Cybernetica, 20:167-187, 2011. [PDF]
    Abstract: It has already been shown that the choice of projection angles can significantly influence the quality of reconstructions in discrete tomography. In this contribution we summarize and extend the previous results by explaining and demonstrating the effects of projection selection dependency, in a set of experimental software tests. We perform reconstructions of software phantoms, by using different binary tomography reconstruction algorithms, from different equiangular and non-equiangular projections sets, under various conditions (i.e., when the objects to be reconstructed undergo slight topological changes, or the projection data is affected by noise) and compare the results with suitable approaches. Based on our observations, we reveal regularities in the resulting data and discuss possible consequences of such projection selection dependency in binary tomography. [bibtex-key = VargaBalazsNagy2011Acta]


16. Rudiger Bock, Jorg Meier, Laszlo G. Nyul, Joachim Hornegger, and Georg Michelson. Glaucoma Risk Index: Automated glaucoma detection from color fundus images. Medical Image Analysis, 14(3):471-481, 2010. [PDF] [doi:10.1016/j.media.2009.12.006]
    Abstract: Glaucoma as a neurodegeneration of the optic nerve is one of the most common causes of blindness. Because revitalization of the degenerated nerve fibers of the optic nerve is impossible early detection of the disease is essential. This can be supported by a robust and automated mass-screening. We propose a novel automated glaucoma detection system that operates on inexpensive to acquire and widely used digital color fundus images. After a glaucoma specific preprocessing, different generic feature types are compressed by an appearance-based dimension reduction technique. Subsequently, a probabilistic two-stage classification scheme combines these features types to extract the novel Glaucoma Risk Index (GRI) that shows a reasonable glaucoma detection performance. On a sample set of 575 fundus images a classification accuracy of 80% has been achieved in a 5-fold cross-validation setup. The GRI gains a competitive area under ROC (AUC) of 88% compared to the established topography-based glaucoma probability score of scanning laser tomography with AUC of 87%. The proposed color fundus image-based GRI achieves a competitive and reliable detection performance on a low-priced modality by the statistical analysis of entire images of the optic nerve head. [bibtex-key = Bock:2010:AGD]


17. Csaba Domokos and Zoltan Kato. Parametric Estimation of Affine Deformations of Planar Shapes. Pattern Recognition, 43(3):569-578, March 2010. [PDF] [doi:10.1016/j.patcog.2009.08.013]
    Abstract: We consider the estimation of affine transformations aligning a known 2D shape and its distorted observation. The classical way to solve this registration problem is to find correspondences between the shapes and then compute the transformation parameters from these landmarks. Here we propose a novel approach where the exact transformation is obtained as the solution of a polynomial system of equations. The method has been tested on synthetic as well as on real images and its robustness in the presence of segmentation errors and additive geometric noise has also been demonstrated. We have successfully applied the method for the registration of hip prosthesis X-ray images. The advantage of the proposed solution is that it is fast, easy to implement, has linear time complexity, works without established correspondences and provides an exact solution regardless of the magnitude of transformation. [bibtex-key = Domokos-Kato2009a]


18. Endre Varga and Balazs Erdohelyi. Severe Pelvic Bleeding: The Role of Primary Internal Fixation. European Journal of Trauma and Emergency Surgery, 36(2):107-116, 2010. [PDF] [doi:10.1007/s00068-010-1003-7]
    Abstract: Pelvic ring injuries form part of the spectrum of polytrauma and must be considered a potentially lethal injury with mortality rates of 10-20%. The stabilization of the unstable pelvic ring in acute resuscitation of multiply injured patients is now conventional wisdom. We aimed: (1) to design a new iliosacral (IS) screw, (2) to prove the clinical advantages of this new implant, and (3) to work out the optimal surgical strategy using this implant. Taking the demands of the above mentioned data into account, a 10 mm 2.8 mm-cannulated iliosacral screw seemed to be optimal for the special requirements. Before industrial production, finite element analysis (FEA) was performed to find out whether these screws would be enough to stabilize the posterior pelvic ring alone or not. Clinical experience led to the modification of the set of instruments, which finally yielded handy tools and implants. Building further on the surgical skills and experiences gained (by the surgeons and the O.R. personnel), we increased our capacity to perform more and more immediate pelvic fixations. Emergency pelvic stabilizationswere performed in patientswith pelvic injuries who had hemodynamic instability, despite immediate shockmanagement during the diagnostic period. During thelast eleven years, 244 patientswith Tile B3 and Cpelvic injuries have been stabilized with 10-mm diameter cannulated IS screws percutaneously posteriorly. Fortyeight hemodynamically unstable patientswere stabilized in the first 2 h with iliosacral screw fixation. The percutaneous pelvic ring stabilization with 10-mm cannulated screws provedstrongenoughin bothersome casesaswell. [bibtex-key = VargaEJTES2010]


19. Gabor Nemeth and Kalman Palagyi. Parallel Thinning Algorithms Based on Ronse's Sufficient Conditions for Topology Preservation. In Progress in Combinatorial Image Analysis, pages 183-194. Research Publishing, 2009. [PDF]
    Abstract: Thinning is a widely used pre?processing step in digital image processing and pattern recognition. It is an iterative layer by layer erosion until only the "skeletons" of the objects are left. This paper presents three thinning algorithms according to three kinds of endpoint criteria. The strategy which is used is called fully parallel, which means that the same parallel operator is applied at each iteration. The proposed algorithms are based on Ronse?s sufficient conditions for topology preservation. [bibtex-key = NemethPalagyi09]


20. Peter Balazs. A benchmark set for the reconstruction of hv-convex discrete sets. Discrete Applied Mathematics, 157(16):3446-3456, 2009. Note: Combinatorial Approach to Image Analysis. [WWW] [PDF] [doi:10.1016/j.dam.2009.02.019]
    Abstract: In this paper we summarize the most important generation methods developed for the subclasses of hv-convex discrete sets. We also present some new generation techniques to complement the former ones thus making it possible to design a complete benchmark set for testing the performance of reconstruction algorithms on the class of hv-convex discrete sets and its subclasses. By using this benchmark set the paper also collects several statistics on hv-convex discrete sets, which are of great importance in the analysis of algorithms for reconstructing such kinds of discrete sets. [bibtex-key = Balazs2009dam]


21. Csaba Benedek, Tamas Sziranyi, Zoltan Kato, and Josiane Zerubia. Detection of Object Motion Regions in Aerial Image Pairs with a Multilayer Markovian Model. IEEE Transactions on Image Processing, 18(10):2303-2315, October 2009. [PDF] [doi:10.1109/TIP.2009.2025808]
    Abstract: We propose a new Bayesian method for detecting the regions of object displacements in aerial image pairs. We use a robust but coarse 2D image registration algorithm. Our main challenge is to eliminate the registration errors from the extracted change map. We introduce a three-layer Markov random field (L^3MRF) model which integrates information from two different features, and ensures connected homogenous regions in the segmented images. Validation is given on real aerial photos. [bibtex-key = Benedek-etal2009]


22. Mihaly Gara, Tamas Samuel Tasi, and Peter Balazs. Learning connectedness and convexity of binary images from their projections. Pure Mathematics and Applications, 20:1-2:27-48, 2009. [PDF]
    Abstract: In this paper we investigate the retrieval of geometrical information (especially, convexity and connectedness) of binary images from their projections which can be useful in binary tomography to facilitate the task of reconstruction. Supposing that the projections are the features of the images, we study how decision trees, neural networks, and nearest neighbor learning algorithms perform in classifying binary images with different connectedness and convexity properties. [bibtex-key = GaraTasiBalazs2009]


23. Peter Horvath, Ian Jermyn, Zoltan Kato, and Josiane Zerubia. A higher-order active contour model of a `gas of circles' and its application to tree crown extraction. Pattern Recognition, 42(5):699-709, May 2009. [PDF] [doi:10.1016/j.patcog.2008.09.008]
    Abstract: We present a model of a `gas of circles': regions in the image domain composed of a unknown number of circles of approximately the same radius. The model has applications to medical, biological, nanotechnological, and remote sensing imaging. The model is constructed using higher-order active contours (HOACs) in order to include non-trivial prior knowledge about region shape without constraining topology. The main theoretical contribution is an analysis of the local minima of the HOAC energy that allows us to guarantee stable circles, fix one of the model parameters, and constrain the rest. We apply the model to tree crown extraction from aerial images of plantations. Numerical experiments both confirm the theoretical analysis and show the empirical importance of the prior shape information. [bibtex-key = Horvath-etal2009]


24. Antal Nagy. Emisszios Diszkret Tomografiai Modszerek Alkalmazasa Faktorstrukturakra. Alkalmazott Matematikai Lapok, 26:329-350, 2009. [PDF]
    Abstract: Az utobbi evekben egy uj fajta diszkret tomografiai problema kutatasa kezdodott el [3], amit emisszios diszkret tomografianak, roviden EDT-nek nevezunk. Ebben a modellben a teljes ter valamilyen homogen abszorbens anyaggal van kitoltve es a rekonstrualando fuggveny egy targyat reprezent al, aminek a pontjai (radioaktiv) sugarzast bocsatanak ki a kornyezo terbe. A targy egy pontjabol kibocsatott kezdeti aktivitas egy resze az abszorbens anyagban elnyelodik a pont es a detektor tavolsagatol fuggoen. A rekonstrukcio kiindulasara szolgalo vetuletek tehat nem tisztan az emissziora vonatkozo adatokat tartalmazzak, hanem az abszorpcio hatasat is. A fo celunk az volt, hogy olyan strukturak terfogatat becsuljuk meg nehany (jelen esetben 4) vetuletbol, melyeknek intenzitasa az idoben valtozik. Mindegyik vetulet adott ido alatt lett elkeszitve, melyek az adott pillanatban felvett kepek sorozatabol allt. A strukturak vetuletei eloszor faktoranalizissel lettek elkulonitve a teljes vetuleti adatsorozatot felhasznalva. Mivel a faktoranalizis a strukturak vetuleteit csak egy szorzo konstans [4] erejeig kepes meghatarozni, ezert a korabbi heurisztikus modszer mellett egy uj modszert is bemutatunk, mely az abszorpcios vetuletek konzisztencia feltetelen [2, 8] alapszik. Ezek utan a mindegyik strukturat kulon-kulon rekonstrualtuk az adott szorzo konstansokkal modositott faktor vetuletekbŽl diszkret tomografiai modszerrel. Az igy kapott terfogatokat az adott strukturak eseten osszehasonlitottuk a cikkben az adott szorzo konstansokat meghatarozo modszerek eseten. [bibtex-key = NagyA_AlkmatLapok2009]


25. Erich Sorantin, Emese Balogh, Anna Vilanova i Bartroli, Kalman Palagyi, Laszlo G. Nyul, Franz Lindbichler, and Andrea Ruppert. Techniques of Virtual Dissection of the Colon Based on Spiral CT Data. In Emanuele Neri, Davide Caramella, and Carlo Bartolozzi, editors, Image Processing in Radiology: Current Applications, Medical Radiology, chapter 18, pages 257-268. Springer Verlag, Berlin, Heidelberg, 2008. [doi:10.1007/978-3-540-49830-8_18]
    Abstract: Colorectal cancer represents the third most commonly diagnosed cancer and is the second leading cause of cancer deaths in the United States (Gazelle et al. 2000). In addition, colorectal cancer is responsible for about 11% of all new cancer cases per year (Gazelle et al. 2000). Five-year prognosis is about 90% for patients with localized disease compared to 60% if there is a regional spread and a drop to 10% in patients with distant metastasis (Gazelle et al. 2000). In the field of medicine there is a widely accepted opinion that most colorectal cancers arise from pre-existent adenomatous polyps (Johnson 2000). Therefore, different societies, such as the American Cancer Society, have proposed screening for colorectal cancer (Byers et al. 1997; Winawer et al. 1997). Today, different options exist for detection of colorectal cancer, including digital rectal examination, fecal occult blood testing, flexible and rigid sigmoidoscopy, barium enema and its variants, colonoscopy and recently computed tomography or magnetic resonance-based virtual colonography (Gazelle et al. 2000). [bibtex-key = SorantinIPR2008]


26. Peter Balazs. A framework for generating some discrete sets with disjoint components by using uniform distributions. Theoretical Computer Science, 406:15-23, 2008. [PDF]
    Abstract: Discrete tomography deals with the reconstruction of discrete sets from few projections. Assuming that the set to be reconstructed belongs to a certain class of discrete sets with some geometrical properties is a commonly used technique to reduce the number of possibly many different solutions of the same reconstruction problem. The average performance of reconstruction algorithms are often tested on such classes by choosing elements of a given class from uniform random distributions. This paper presents a general framework for generating discrete sets with disjoint connected components using uniform distributions. Especially, the uniform random generation of $hv$-convex discrete sets and Q-convex discrete sets according to the size of the minimal bounding rectangle are discussed. [bibtex-key = Balazs2008d]


27. Peter Balazs. Discrete tomographic reconstruction of binary images with disjoint components using shape information. International Journal of Shape Modeling, 14:2:189-207, 2008. [PDF]
    Abstract: We present a general framework for reconstructing binary images with disjoint components from the horizontal and vertical projections. We develop a backtracking algorithm that works for binary images having components from an arbitrary class. Thus, a priori knowledge about the components of the image to be reconstructed can be incorporated into the reconstruction process. In addition, we show how to extend the algorithm to obtain a branch-and-bound scheme useful to reconstruct images satisfying some further properties (for example similarity to a model image) as much as possible. Experimental results are also presented. [bibtex-key = Balazs2008ws]


28. Peter Balazs. On the Ambiguity of Reconstructing hv-Convex Binary Matrices with Decomposable Configurations. Acta Cybernetica, 18(3):367-377, 2008. [PDF] Keyword(s): discrete tomography, hv-convex binary matrix, decomposable configuration, reconstruction algorithm.
    Abstract: Reconstructing binary matrices from their row, column, diagonal, and antidiagonal sums (also called projections) plays a central role in discrete tomography. One of the main difficulties in this task is that in certain cases the projections do not uniquely determine the binary matrix. This can yield an extremely large number of (sometimes very different) solutions. This ambiguity can be reduced by having some prior knowledge about the matrix to be reconstructed. The main challenge here is to find classes of binary matrices where ambiguity is drastically reduced or even completely eliminated. The goal of this paper is to study the class of $hv$-convex matrices which have decomposable configurations from the viewpoint of ambiguity. First, we give a negative result in the case of three projections. Then, we present a heuristic for the reconstruction using four projections and analyze its performance in quality and running time. [bibtex-key = Balazs:2008:ActaCybernetica]


29. Gyorgy Bekes, Eors Mate, Laszlo G. Nyul, Attila Kuba, and Marta Fidrich. Geometrical model-based segmentation of the organs of sight on CT images. Medical Physics, 35(2):735-743, February 2008. [WWW] [PDF] [doi:10.1118/1.2826557] Keyword(s): biological tissues, computerised tomography, eye, image segmentation, medical image processing, neurophysiology, radiation therapy.
    Abstract: Segmentation of organs of sight such as the eyeballs, lenses, and optic nerves is a time consuming task for clinicians. The small size of the organs and the similar density of the surrounding tissues make the segmentation difficult. We developed a new algorithm to segment these organs with minimal user interaction. The algorithm needs only three seed points to fit an initial geometrical model to start an effective segmentation. The clinical evaluation shows that the output of our method is useful in clinical practice. [bibtex-key = bekes:735]


30. Zoltan Kato. Segmentation of Color Images via Reversible Jump MCMC Sampling. Image and Vision Computing, 26(3):361-371, March 2008. [PDF] [doi:10.1016/j.imavis.2006.12.004]
    Abstract: Reversible jump Markov chain Monte Carlo (RJMCMC) is a recent method which makes it possible to construct reversible Markov chain samplers that jump between parameter subspaces of different dimensionality. In this paper, we propose a new RJMCMC sampler for multivariate Gaussian mixture identification and we apply it to color image segmentation. For this purpose, we consider a first order Markov random field (MRF) model where the singleton energies derive from a multivariate Gaussian distribution and second order potentials favor similar classes in neighboring pixels. The proposed algorithm finds the most likely number of classes, their associated model parameters and generates a segmentation of the image by classifying the pixels into these classes. The estimation is done according to the Maximum A Posteriori (MAP) criterion. The algorithm has been validated on a database of real images with human segmented ground truth. [bibtex-key = Kato2007]


31. Kalman Palagyi. A 3D fully parallel surface-thinning algorithm. Theoretical Computer Science, 406:119-135, 2008. [PDF] [doi:10.1016/j.tcs.2008.06.041]
    Abstract: The thinning is an iterative layer by layer erosion until only the ``skeletons'' of the objects are left. This paper presents a thinning algorithm for extracting medial surfaces from 3D binary pictures. The strategy which is used is called fully parallel, which means that the same parallel operator is applied at each iteration. An efficient implementation of the proposed algorithm on conventional sequential computers is given and the topological correctness for (26, 6) binary pictures is proved. [bibtex-key = PalagyiTCS2008]


32. Peter Balazs. Decomposition algorithms for reconstructing discrete sets with disjoint components. In Gabor T. Herman and Attila Kuba, editors, Advances in Discrete Tomography and Its Applications, chapter 8, pages 153-173. Birkhauser, Boston, 2007. [WWW] [PDF]
    Abstract: The reconstruction of discrete sets from their projections is a frequently studied field in discrete tomography with applications in electron microscopy, image processing, radiology, and so on. There have been several efficient reconstruction algorithms developed for certain classes of discrete sets having some good geometrical properties. On the other side it has been shown that the reconstruction under certain circumstances can be very time-consuming, or even NP-hard. In this chapter we show how the prior information that the set to be reconstructed consists of several components can be exploited in order to facilitate the reconstruction. We present some general techniques to decompose a discrete set into single components knowing only its projections thus reducing the reconstruction of a general discrete set to the reconstruction of single components which is usually a simpler task. [bibtex-key = Balazs2007b]


33. Elena Barcucci, Andrea Frosini, Attila Kuba, Antal Nagy, Simone Rinaldi, Martin Samal, and Steffen Zopf. Emission discrete tomography. In Gabor T. Herman and Attila Kuba, editors, Advances in Discrete Tomography and Its Applications, pages 333-366. Birkhauser, Boston, 2007. [WWW] [PDF]
    Abstract: Three problems of emission discrete tomography (EDT) are presented. The first problem is the reconstruction of measurable plane sets from two absorbed projections. It is shown that Lorentz theorems can be generalized to this case. The second is the reconstruction of binary matrices from their absorbed row and column sums if the absorption coefficient is $\mu_{0} = log((1+\sqrt{5})/2)$. It is proved that the reconstruction in this case can be done in polynomial time. Finally, a possible application of EDT in single photon emission computed tomography (SPECT) is presented: Dynamic structures are reconstructed after factor analysis. [bibtex-key = BarcucciFrosiniKubaNagyRinaldiSamalZopf:2007:EDT]


34. Joachim Baumann, Zoltan Kiss, Sven Krimmel, Attila Kuba, Antal Nagy, Lajos Rodek, Burkhard Schillinger, and Juergen Stephan. Discrete Tomography Methods for Nondestructive Testing. In Gabor T. Herman and Attila Kuba, editors, Advances in Discrete Tomography and Its Applications, pages 303-332. Birkhauser, Boston, 2007. [WWW] [PDF]
    Abstract: The industrial nondestructive testing (NDT) of objects seems to be an ideal application of discrete tomography. In many cases, the objects consist of known materials, and a lot of a prior infromation is available (e.g., the description of an ideal object, which is similar to the actual one under investigation. One of frequently used methods in NDT is to take projection images of the objects by some transmitting ray (e.g., X- or neutron ray) and reconstruct the cross sections. But it can happen that only a few number of projections can be collected, because of long and/or expensive data acquisition, or the projection can be collected only from a limited range of direction. The chapter describes two DT reconstruction methods used in NDT experiments, shows the results of a DT procedure applied in the reconstruction of oblong objects having projections only from a limited range of angles, and, finally suggests a few further possible NDT applications of DT. [bibtex-key = BaumannKissKrimmelKubaNagyRodekSchillingerStephan:2007:NDT]


35. Peter Balazs. A decomposition technique for reconstructing discrete sets from four projections. Image and Vision Computing, 25:10:1609-1619, 2007. [PDF]
    Abstract: The reconstruction of discrete sets from four projections is in general an NP-hard problem. In this paper we study the class of decomposable discrete sets and give an efficient reconstruction algorithm for this class using four projections. It is also shown that an arbitrary discrete set which is Q-convex along the horizontal and vertical directions and consists of several components is decomposable. As a consequence of decomposability we get that in a subclass of $hv$-convex discrete sets the reconstruction from four projections can also be solved in polynomial time. Possible extensions of our method are also discussed. [bibtex-key = Balazs2007a]


36. Endre Katona. Contour line thinning and multigrid generation of raster-based digital elevation models. International Journal of Geographical Information Science, 21:71-82, 2007. [PDF]
    Abstract: Thin plate spline interpolation is a widely used approach to generate a digital elevation model (DEM) from contour lines and scattered data. In practice, contour maps are scanned and vectorized, and after resampling in the target grid resolution, interpolation is performed. In this paper we demonstrate the limited accuracy of this process, and propose a high resolution processing method (without vectorization) that ensures maximum utilization of information in the source data. First, we discuss the mathematical background of thin plate spline interpolation, and explain the multigrid relaxation principle used to speed up convergence. After, we will show why fine tuning is necessary, especially when contour lines and elevation points are processed at the same time. Finally, our own contour thinning method that produces a significant reduction of elevation bias is described. [bibtex-key = Katona2007]


37. Gabor Fichtinger, Everette C. Burdette, Attila Tanacs, Alexandru Patriciu, Dumitru Mazilu, Louis L. Whitcomb, and Dan Stoianovici. Robotically assisted prostate brachytherapy with transrectal ultrasound guidance --- Phantom experiments. Brachytherapy, 5:14-26, 2006. [PDF] [doi:doi:10.1016/j.brachy.2005.10.003]
    Abstract: {\bf PURPOSE:} To report the preliminary experimental results obtained with a robot-assisted trans-rectal ultrasound (TRUS)--guided prostate brachytherapy system. {\bf METHODS AND MATERIALS:} The system consists of a TRUS unit, a spatially coregistered needle insertion robot, and an FDA-approved treatment planning and image-registered implant system. The robot receives each entry/target coordinate pair of the implant plan, inserts a preloaded needle, and then the seeds are deposited. The needles/sources are tracked in TRUS, thus allowing the plan to be updated as the procedure progresses. {\bf RESULTS:} The first insertion attempt was recorded for each needle, without adjustment. All clinically relevant locations were reached in a prostate phantom. Nonparallel and parallel needle trajectories were demonstrated. Based on TRUS, the average transverse placement error was 2 mm (worst case 2.5 mm, 80% less than 2 mm), and the average sagittal error was 2.5 mm (worst case 5.0 mm, 70% less than 2.5 mm). {\bf CONCLUSIONS:} The concept and technical viability of robot-assisted brachytherapy were demonstrated in phantoms. The kinematically decoupled robotic assistant device is inherently safe. Overall performance was promising, but further optimization is necessary to prove the possibility of improved dosimetry. [bibtex-key = Fichtinger:2006:Brachytherapy]


38. Zoltan Kato and Ting Chuen Pong. A Markov Random Field Image Segmentation Model for Color Textured Images. Image and Vision Computing, 24(10):1103-1114, October 2006. [PDF]
    Abstract: We propose a Markov random field (MRF) image segmentation model which aims at combining color and texture features. The theoretical framework relies on Bayesian estimation via combinatorial optimization (Simulated Annealing). The segmentation is obtained by classifying the pixels into different pixel classes. These classes are represented by multi-variate Gaussian distributions. Thus, the only hypothesis about the nature of the features is that an additive Gaussian noise model is suitable to describe the feature distribution belonging to a given class. Here, we use the perceptually uniform CIE-L*u*v* color values as color features and a set of Gabor filters as texture features. Gaussian parameters are either computed using a training data set or estimated from the input image. We also propose a parameter estimation method using the EM algorithm. Experimental results are provided to illustrate the performance of our method on both synthetic and natural color images. [bibtex-key = Kato-Pong2006a]


39. Zoltan Kiss, Lajos Rodek, and Attila Kuba. Image reconstruction and correction methods in neutron and X-ray tomography. Acta Cybernetica, 17(3):557-587, 2006. [WWW]
    Abstract: Neutron and X-ray tomography are imaging techniques for getting information about the interior of objects in a non-destructive way. They reconstruct cross-sections from projection images of the object being investigated. Due to the properties of the image acquisition system, the projection images are distorted by several artifacts, and these reduce the quality of the reconstruction. In order to eliminate these harmful effects the projection images should be corrected before reconstruction. Taking projections is usually an expensive and time consuming procedure. One of our main goals has been to try to minimize the number of projections - for example, by exploiting more a priori information. A possible way of reducing the number of projections is by the application of discrete tomographic methods. In this case a special class of objects can be reconstructed, consisting of only a few homogenous materials that can be characterized by known discrete absorption values. To this end we have implemented two reconstruction methods. One is able to reconstruct objects consisting of cylinders and spheres made of homogeneous materials only. The other method is a general one in the sense that it can be used for reconstructing any shape. Simulations on phantoms and physical measurements were carried out and the results are presented here. [bibtex-key = Kiss2006356]


40. Antal Nagy and Attila Kuba. Parameter settings for reconstructing binary matrices from fan-beam projections. Journal of Computing and Information Technology, 14(2):100-110, 2006. [PDF]
    Abstract: The problem of reconstruction of binary matrices from their fan-beam projections is studied. A fan-beam pro- jection model is implemented and used in systematic ex- periments in order to determine the optimal parameter values for data acquisition and reconstruction algorithm. The fan-beam model, the reconstruction algorithm, the simulation experiments, and the results are discussed in the paper. [bibtex-key = NagyKuba:2004:ParamFanBeam]


41. Krisztian Olle, Balazs Erdohelyi, Endre Varga, Csongor Halmai, and Attila Kuba. MedEdit: A Computer Assisted Image Processing and Navigation System for Orthopedic Trauma Surgery. Acta Cybernetica, 17:589-603, July 2006. [PDF]
    Abstract: The surgery of fractured bones is often a very complex problem. That is the reason why it would be beneficial to create a geometric and mechanic model of the bones before surgical intervention. The model geometry is based on the CT images of the patient and the known physical properties of the bone. A computerised system is presented here, called MedEdit, which helps a surgeon plan an operation. The system includes a Finite Element Analysis (FEA) program to measure the stress effects of the possible surgical solutions. Following the simulation and analysis of the behaviour of the modelled bone, surgeons can find the best surgical solution for the patient. [bibtex-key = Olle2006]


42. Kalman Palagyi, Juerg Tschirren, Eric A. Hoffman, and Milan Sonka. Quantitative analysis of pulmonary airway tree structure. Computers in Biology and Medicine, 36:974-996, 2006. [PDF]
    Abstract: A method for computationally efficient skeletonization of three-dimensional tubular structures is reported. The method is specifically targeting skeletonization of vascular and airway tree structures in medical images but it is general and applicable to many other skeletonization tasks. The developed approach builds on the following novel concepts and properties: fast curve-thinning algorithm to increase computational speed, endpoint re-checking to avoid generation of spurious side branches, depth-and-length sensitive pruning, and exact tree-branch partitioning allowing branch volume and surface measurements. The method was validated in computer and physical phantoms and in vivo CT scans of human lungs. The validation studies demonstrated sub-voxel accuracy of branch point positioning, insensitivity to changes of object orientation, and high reproducibility of derived quantitative indices of the tubular structures offering a significant improvement over previously reported methods (p>>0.001). [bibtex-key = PalagyiEtalCBM2006]


43. Peter Balazs, Emese Balogh, and Attila Kuba. Reconstruction of 8-connected but not 4-connected hv-convex discrete sets. Discrete Applied Mathematics, 147:149-168, 2005. [PDF]
    Abstract: The reconstruction of 8-connected but not 4-connected $hv$-convex discrete sets from few projections is considered. An algorithm is given with worst case complexity of $O(mn\cdot\min\{ m,n\} )$ to reconstruct all sets with given horizontal and vertical projections. Experimental results are also presented. It is shown, that using also the diagonal projections the algorithm can be speeded up having complexity of $O(mn)$ and in this case the solution is uniquely determined. Finally, we consider the possible generalizations of our results to solve the problem in more general classes. [bibtex-key = Peter2005]


44. Attila Kuba and Maurice Nivat. A sufficient condition for non-uniqueness in binary tomography with absorption. Theoretical Computer Science, 346:335-357, November 2005. [PDF] [doi:10.1016/j.tcs.2005.08.024]
    Abstract: A new kind of discrete tomography problem is introduced: the reconstruction of discrete sets from their absorbed projections. A special case of this problem is discussed, namely, the uniqueness of the binary matrices with respect to their absorbed row and column sums when the absorption coefficient is n=log((1+5^0.5)/2). It is proved that if a binary matrix contains a special structure of 0s and 1s, called alternatively corner-connected component, then this binary matrix is non-unique with respect to its absorbed row and column sums. Since it has been proved in another paper [A. Kuba, M. Nivat, Reconstruction of discrete sets with absorption, Linear Algebra Appl. 339 (2001) 171–194] that this condition is also necessary, the existence of alternatively corner-connected component in a binary matrix gives a characterization of the non-uniqueness in this case of absorbed projections. [bibtex-key = Kuba2005]


45. Attila Kuba, Laszlo Rusko, Lajos Rodek, and Zoltan Kiss. Preliminary studies of discrete tomography in neutron imaging. IEEE Transactions on Nuclear Science, 52(1):380-385, February 2005. [doi:10.1109/TNS.2005.843657]
    Abstract: Discrete tomography (DT) is a new technique to reconstruct discrete images from their projections (like neutron images). The reconstruction methods in DT are different from the conventional ones, because the created images may contain only a few numbers of given discrete values. One of the main reasons to apply DT is that hopefully we need only a few numbers of projections. In many applications we have a situation where we know the material components of the object to be studied, that is, we know the discrete values of the image to be reconstructed. Using discreteness and some a priori information we can apply several DT methods in neutron imaging. Most of the DT reconstruction methods are reducing the problem to an optimization task. We tried two such methods on software and physical phantoms. In these experiments we investigated the effects of the following parameters: number of projections, noise levels, and complexity of the object to be reconstructed. We also developed a software system, called DIRECT, for testing different DT methods, to compare them and to present the reconstructed objects. [bibtex-key = KubaRusko200552]


46. Antal Nagy and Attila Kuba. Reconstruction of binary matrices from fan-beam projections. Acta Cybernetica, 17(2):359-385, 2005. [PDF]
    Abstract: The problem of the reconstruction of binary matrices from their fan-beam projections is investigated here. A fan-beam projection model is implemented and afterwards employed in systematic experiments to determine the optimal parameter values for a data acquisition and reconstruction algorithm. The fan-beam model, the reconstruction algorithm which uses the optimization method of Simulated Annealing, the simulation experiments, and the results are then discussed in turn. [bibtex-key = NagyKuba:2005:RecFanBeam]


47. Juerg Tschirren, Geoffrey McLennan, Kalman Palagyi, Eric A. Hoffman, and Milan Sonka. Matching and anatomical labeling of human airway tree. IEEE Transactions on Medical Imaging, 24:1540-1547, 2005. [PDF]
    Abstract: Matching of corresponding branchpoints between two human airway trees, as well as assigning anatomical names to the segments and branchpoints of the human airway tree, are of significant interest for clinical applications and physiological studies. In the past, these tasks were often performed manually due to the lack of automated algorithms that can tolerate false branches and anatomical variability typical for in vivo trees. In this paper, we present algorithms that perform both matching of branchpoints and anatomical labeling of in vivo trees without any human intervention and within a short computing time. No hand-pruning of false branches is required. The results from the automated methods show a high degree of accuracy when validated against reference data provided by human experts. 92.9% of the verifiable branchpoint matches found by the computer agree with experts’ results. For anatomical labeling, 97.1% of the automatically assigned segment labels were found to be correct. [bibtex-key = TschirrenEtalTMI2005]


48. Attila Kuba, Antal Nagy, and Emese Balogh. Reconstruction of hv-convex binary matrices from their absorbed projections. Discrete Applied Mathematics, 139:137-148, April 2004. [PDF] [doi:10.1016/j.dam.2002.11.001]
    Abstract: The reconstruction of hv-convex binary matrices from their absorbed projections is considered. Although this problem is NP-hard if the non-absorbed row and column sums are available, it is proved that such a reconstruction problem can be solved in polynomial time from absorbed projections when the absorption is represented by $\beta = (1+sqrt{5})/2$. Also a reconstruction algorithm is given to determine the whole structure of hv-convex binary matrices from such projections. [bibtex-key = Kuba2004]


49. Burkhard Schillinger, N. Kardjilov, and Attila Kuba. Region of interest tomography of bigger than detector samples. Applied Radiation and Isotopes, 61(4):561-565, October 2004. [PDF] [doi:10.1016/j.apradiso.2004.03.092]
    Abstract: For many neutron tomography setups, the maximum sample size for tomography is limited by a comparatively small beam cross section. However, it is not well known outside the medical field that it is possible to perform region-of-interest tomography of sections inside the object that fit into the beam and detector area. Approximately valid reconstruction data appear in a circle with a diameter of approximately the detector width, but with incomplete data and strong artifacts outside that circle. These artifacts can be removed either by mathematical means or by simple geometrical cutting of the reconstructed data, enabling the examination of samples bigger than the detector or beam area. [bibtex-key = Schillinger2004]


50. Attila Tanacs. Kijelolt pontparokon alapulo kepregisztracios modszerek. Alkalmazott Matematikai Lapok, 21:237-260, 2004. [PDF]
    Abstract: Registration is a fundamental task in digital image processing. Its purpose is to find a geometrical transformation that relates the points of an image to their corresponding points of another image. A general and easy to use solution for registration problems is selecting pairs of points as features. In this paper, after a short general introduction which describes the most common features of the registration methods, we discuss some of the point based methods proposed by us and other authors. We successfully use these in the field of medical image registration but they also can be used in other fields including computer vision and remotely sensed data processing. We didn't inted to describe all the available methods. For some transformation types such as rigid-body, there are several others with similar complexity and reliability. We intended to give at least one useful method for every transformation types. [bibtex-key = Tanacs:2004:AlkMatLapok]


51. Gabor T. Herman and Attila Kuba. Discrete tomography in medical imaging. Proceedings of the IEEE, 91:1612-1626, October 2003. [doi:10.1109/JPROC.2003.817871]
    Abstract: Discrete tomography (DT) deals with the reconstruction of a function from its projections, when the function has a known discrete range. The knowledge of the discrete range, possibly together with some prior information, can significantly reduce the number of projections required for a high-quality reconstruction. The reconstruction methods used in DT applications are usually based on some formulation as an optimization problem. This paper presents methods and results of DT based on problems of angiography, emission tomography, and electron microscopy (EM). [bibtex-key = GTHerman2003ProcIeee46]


52. Eric A. Hoffman, Joseph M. Reinhardt, Milan Sonka, Brett A. Simon, Junfeng Guo, Osama Saba, Deokiee Chon, Shaher Samrah, Hidenori Shikata, Juerg Tschirren, Kalman Palagyi, Kenneth C. Beck, and Geoffrey McLennan. Characterization of the Interstitial Lung Diseases via Density-Based and Texture-Based Analysis of Computed Tomography Images of Lung Structure and Function. Academic Radiology, 10:1104-1118, 2003. [PDF]
    Abstract: Rationale and Objectives. Efforts to establish a quantitative approach to the computed tomography (CT)-based characterization of the lung parenchyma in interstitial lung disease (including emphysema) has been sought. The accuracy of these tools must be site independent. Multi-detector row CT has remained the gold standard for imaging the lung, and it provides the ability to image both lung structure as well as lung function. Material and Methods. Imaging is via multi-detector row CT and protocols include careful control of lung volume during scanning. Characterization includes not only anatomic-based measures but also functional measures including regional parameters derived from measures of pulmonary blood flow and ventilation. Image processing includes the automated detection of the lungs, lobes, and airways. The airways provide the road map to the lung parenchyma. Software automatically detects the airways, the airway centerlines, and the branch points, and then automatically labels the airway tree segments with a standardized set of labels, allowing for intersubject as well intrasubject comparisons across time. By warping all lungs to a common atlas, the atlas provides the range of normality for the various parameters provided by CT imaging. Results. Imaged density and textural changes mark underlying structural changes at the most peripheral regions of the lung. Additionally, texture-based alterations in the parameters of blood flow may provide early evidence of pathologic processes. Imaging of stable xenon gas provides a regional measure of ventilation which, when coupled with measures of flow, provide for a textural analysis regional of ventilation-perfusion matching. Conclusion. With the improved resolution and speed of CT imaging, the patchy nature of regional parenchymal pathology can be imaged as texture of structure and function. With careful control of imaging protocols and the use of objective image analysis methods it is possible to provide site-independent tools for the assessment of interstitial lung disease. There remains a need to validate these methods, which requires interdisciplinary and cross-institutional efforts to gather appropriate data bases of images along with a consensus on appropriate ground truths associated with the images. Furthermore, there is the growing need for scanner manufacturers to focus on not just visually pleasing images, but on quantitatifiably accurate images. Key Words. Quantitative CT; airways; parenchyma; emphysema; interstitial lung disease; computer analysis; blood flow; ventitation. [bibtex-key = HoffmanEtalAcadRad2003]


53. Tianhu Lei, Jayaram K. Udupa, Dewey Odhner, Laszlo G. Nyul, and Punam K. Saha. 3DVIEWNIX-AVS: A Software Package for the Separate Visualization of Arteries and Veins in CE-MRA Images. Computerized Medical Imaging and Graphics, 27(5):351-362, September 2003. [PDF] [doi:10.1016/S0895-6111(03)00029-6]
    Abstract: Our earlier study developed a computerized method, based on fuzzy connected object delineation principles and algorithms, for artery and vein separation in contrast enhanced Magnetic Resonance Angiography (CE-MRA) images. This paper reports its current development ``a software package'' for routine clinical use. The software package, termed 3DVIEWNIX-AVS, consists of the following major operational parts: (1) converting data from DICOM3 to 3DVIEWNIX format, (2) previewing slices and creating VOI and MIP Shell, (3) segmenting vessel, (4) separating artery and vein, (5) shell rendering vascular structures and creating animations. This package has been applied to EPIX Medical Inc's CE-MRA data (AngioMark MS-325). One hundred and thirty-five original CE-MRA data sets (of 52 patients) from 6 hospitals have been processed. In all case studies, unified parameter settings produce correct artery-vein separation. The current package is running on a Pentium PC under Linux and the total computation time per study is about 3 min. The strengths of this software package are (1) minimal user interaction, (2) minimal anatomic knowledge requirements on human vascular system, (3) clinically required speed, (4) free entry to any operational stages, (5) reproducible, reliable, high quality of results, and (6) cost effective computer implementation. To date, it seems to be the only software package (using an image processing approach) available for artery and vein separation of the human vascular system for routine use in a clinical setting. [bibtex-key = Lei:2003:AVS]


54. Laszlo G. Nyul, Jayaram K. Udupa, and Punam K. Saha. Incorporating a Measure of Local Scale in Voxel-based 3D Image Registration. IEEE Transactions on Medical Imaging, 22(2):228-237, February 2003. [PDF]
    Abstract: We present a new class of approaches for rigid-body registration and their evaluation in studying multiple sclerosis (MS) via multiprotocol magnetic resonance imaging (MRI). Three pairs of rigid-body registration algorithms were implemented, using cross-correlation and mutual information (MI), operating on original gray-level images, and utilizing the intermediate images resulting from our new scale-based method. In the scale image, every voxel has the local ``scale'' value assigned to it, defined as the radius of the largest ball centered at the voxel with homogeneous intensities. Three-dimensional image data of the head were acquired from tenMSpatients for each of six MRI protocols. Images in some of the protocols were acquired in registration. The registered pairs were used as ground truth. Accuracy and consistency of the six registration methods were measured within and between protocols for known amounts of misregistrations. Our analysis indicates that there is no ``best'' method. For medium misregistration, the method using MI, for small and large misregistration the method using normalized cross-correlation performs best. For high-resolution data the correlation method and for low-resolution data the MI method, both using the original gray-level images, are the most consistent. We have previously demonstrated the use of local scale information in fuzzy connectedness segmentation and image filtering. Scale may also have potential for image registration as suggested by this work. [bibtex-key = Nyul:2003:IML]


55. Erich Sorantin, Csongor Halmai, Balazs Erdohelyi, Kalman Palagyi, Laszlo G. Nyul, Krisztian Olle, Bernhard Geiger, Franz Lindbichler, Gerhard Friedrich, and Karl Kiesler. 3D-Querschnittsprofil des Laryngotrachealtrakts. Der Radiologe, 43(12):1056-1068, December 2003. [WWW] [PDF] [doi:10.1007/s00117-003-0990-8] Keyword(s): Image processing - Biomedical computing - Biomedical measurements - Feature extraction.
    Abstract: Purpose Demonstration of a technique for 3D assessment of tracheal stenoses, regarding site, length and degree, based on spiral computed tomography (S-CT). Patients and Methods S-CT scanning and automated segmentation of the laryngo-tracheal tract (LTT) was followed by the extraction of the LTT medial axis using a skeletonisation algorithm. Orthogonal to the medial axis the LTT 3D cross sectional profile was computed and presented as line charts, where degree and length were obtained. Values for both parameters were compared between 36 patients and 18 normal controls separately. Accuracy and precision was derived from 17 phantom studies. Results Average degree and length of tracheal stenoses were found to be 60.5% and 4.32 cm in patients compared to minor caliber changes of 8.8% and 2.31 cm in normal controls (p <0.005). For the phantoms an excellent correlation between the true and computed 3D cross sectional profile was found (p <0.005) and an accuray for length and degree measurements of 2.14 mm and 2.53% respectively could be determined. The corresponding figures for the precision were found to be 0.92 mm and 2.56%. Conclusion LTT 3D cross sectional profiles permit objective, accurate and precise assessment of LTT caliber changes. Minor LTT caliber changes can be observed even in normals and, in case of an otherwise normal S-CT study, can be regarded as artefacts. [bibtex-key = Sorantin:2003]


56. Robert C. Susil, Axel Krieger, J. Andrew Derbyshire, Attila Tanacs, Louis L. Whitcomb, Gabor Fichtinger, and Ergin Atalar. System for MR Image-guided Prostate Interventions: Canine Study. Journal of Radiology, 228:886-894, 2003. [PDF] [doi:10.1148/radiol.2283020911]
    Abstract: The purpose of this study was to demonstrate the use of a transrectal system that enables precise magnetic resonance (MR) image guidance and monitoring of prostate interventions. The system used a closed-bore 1.5-T MR imaging unit and enables one to take advantage of the higher signal-to-noise ratio achieved with traditional magnet designs, which is crucial for accurate targeting and monitoring of prostate interventions. In the first of the four canine studies, reliable needle placement, with all needles placed within 2 mm of the desired target site, was achieved. In two other studies, MR imaging was used to monitor distribution of injected contrast agent solution (gado-pentetate dimeglumine mixed with trypan blue dye) in and around the prostate, thereby confirming that solution had been delivered to the desired tissue and also detecting faulty injections. In the final study, accurate placement and MR imaging of brachytherapy seeds in the prostate were demonstrated. The described system provides a flexible platform for a variety of minimally invasive MR image--guided therapeutic and diagnostic prostate interventions. [bibtex-key = Susil:2003:Radiology]


57. Attila Tanacs and Attila Kuba. Evaluation of a Fully Automatic Medical Image Registration Algorithm Based on Mutual Information. Acta Cybernetica, 16:327-336, 2003. [PDF]
    Abstract: Registration is a fundamental task in image processing. Its purpose is to find a geometrical transformation that relates the points of an image to their corresponding points of another image. Many registration algorithms have been proposed in the past decade. We present a fast, fully automatic algorithm that is capable of solving rigid-body registration of 3D images of the human brain where the images are taken by different imaging devices. We joined the Retrospective Registration Evaluation Project conducted by Vanderbilt University, USA. The evaluations of our results show that our method has the potential to produce satisfactory results, but visual inspection is necessary to guard against large errors. [bibtex-key = Tanacs:2003:Acta]


58. Erich Sorantin, Emese Balogh, Anna Vilanova i Bartroli, Kalman Palagyi, Laszlo G. Nyul, Sven Loncaric, Marko Subasic, and Domagoj Kovacevic. Virtual Dissection of the Colon. In D. Caramella, editor, 3D Image Processing -- Techniques and Clinical Applications, pages 197-209. Springer Verlag, Heidelberg, February 2002. [bibtex-key = Sorantin:2002:VDC]


59. Marianna Dudasne-Nagy, Eors Mate, and Bela Kari. Mathematical generation of normal data for evaluating myocardial perfusion studies. IEEE Transactions on Medical Imaging, 21:1358- 1365, 2002. [PDF] [doi:10.1109/TMI.2002.806429]
    Abstract: In this paper, we present a new mathematical method that synthesizes normal data sets for quantification of regional myocardium perfusion. In clinical practice, regional myocardial perfusion is often measured with a gamma camera and quantified via circumferential profile analysis. Normal reference profile data is used to increase the accuracy of the clinical interpretations. Our goal is to create reference data from an existing set of archived studies. An iterative mathematical method, based on two statistical hypotheses, was used to generate the study set instead of collecting normal examinations from a healthy population. Clinical validation is based on interpretations by six independent observers. Results of evaluation with synthesized normal data and its validation are presented. [bibtex-key = Dudasne-Nagy2002]


60. Gabor Fichtinger, Theodore L. DeWeese, Alexandru Patriciu, Attila Tanacs, Dumitru Mazilu, James H. Anderson, Ken Masamune, Russel H. Taylor, and Dan Stoianovici. System for Robotically Assisted Prostate Biopsy and Therapy with Intraoperative CT Guidance. Academic Radiology, 9(1):60-74, 2002. [PDF] [doi:10.1016/S1076-6332(03)80297-0]
    Abstract: {\bf Rationale and Objectives}. The purpose of this study was to assess the work-in-progress prototype of an image-guided, robotic system for accurate and consistent placement of transperineal needles into the prostate with intraoperative image guidance inside the gantry of a computed tomographic (CT) scanner. {\bf Materials and Methods.} The coach-mounted system consists of a seven-degrees-of-freedom, passive mounting arm; a remote-center-of-motion robot; and a motorized, radiolucent needle-insertion device to deliver 17--18-gauge implant and biopsy needles into the prostate with the transperineal route. The robot is registered to the image space with a stereotactic adapter. The surgeon plans and controls the intervention in the CT scanner room with a desktop computer that receives DICOM images from the CT scanner. The complete system fits in a carry-on suitcase, does not need calibration, and does not utilize vendor-specific features of the CT scanner. {\bf Results.} In open air, the average accuracy was better than 1 mm at a 5--8-cm depth. In various phantoms, the average orientation error was 1.3 degrees, and the average distance between the needle tip and the target was 2 mm. {\bf Conclusion.} Results of preliminary experiments indicate that this robotic system may be suitable for transperineal needle placement into the prostate and shows potential in a variety of other percutaneous clinical applications. [bibtex-key = Fichtinger:2002:AcadRad]


61. Attila Kuba and Emese Balogh. Reconstruction of convex 2D discrete sets in polynomial time. Theoretical Computer Science, 283:223-242, June 2002. [doi:10.1016/S0304-3975(01)00080-9]
    Abstract: The reconstruction problem is considered in those classes of discrete sets where the reconstruction can be performed from two projections in polynomial time. The reconstruction algorithms and complexity results are summarized in the case of hv-convex sets, hv-convex 8-connected sets, hv-convex polyominoes, and directed h-convex sets. As new results some properties of the feet and spines of the hv-convex 8-connected sets are proven and it is shown that the spine of such a set can be determined from the projections in linear time. Two algorithms are given to reconstruct hv-convex 8-connected sets. Finally, it is shown that the directed h-convex sets are uniquely reconstructible with respect to their row and column sum vectors. [bibtex-key = Kuba2002]


62. Laszlo G. Nyul, Alexandre X. Falcao, and Jayaram K. Udupa. Fuzzy-Connected 3D Object Segmentation at Interactive Speeds. Graphical Models, 64(5):258-281, September 2002. [PDF] [doi:10.1016/S1077-3169(02)00005-9]
    Abstract: Image segmentation techniques using fuzzy connectedness principles have shown their effectiveness in segmenting a variety of objects in several large applications in recent years. However, one problem with these algorithms has been their excessive computational requirements. In an attempt to substantially speed them up, in the present paper, we study systematically a host of 18 'optimal' graph search algorithms. Extensive testing of these algorithms on a variety of 3D medical images taken from large ongoing applications demonstrates that a 20-1000-fold improvement over current speeds is achievable with a combination of algorithms and fast modern PCs. Utilizing efficient algorithms and careful selection of implementations can speed up the computation of fuzzy connectedness values by a factor of 16-29 (on the same hardware), as compared to the implementation previously used in our applications utilizing fuzzy object segmentation. The optimality of an algorithm depends on the input data as well as on the choice of the fuzzy affinity relation. The running time is reduced considerably (by a factor up to 34 for brain MR and even more for bone CT), when the algorithms make use of predetermined thresholds for the fuzzy objects. The reliable recognition (assisted by human operators) and the accurate, efficient, and sophisticated delineation (automatically performed by the computer) can be effectively incorporated into a single interactive process. If images having intensities with tissue-specific meaning (such as CT or standardized MR images) are utilized, most of the parameters for the segmentation method can be fixed once for all, all intermediate data (feature and fuzzy affinity values for the whole scene) can be computed before the user interaction is needed and the user can be provided with more information at the time of interaction. [bibtex-key = Nyul:2002:FC3]


63. Kalman Palagyi. A 3-subiteration 3D thinning algorithm for extracting medial surfaces. Pattern Recognition Letters, 23:663-675, 2002. [PDF]
    Abstract: Thinning is an iterative layer by layer erosion for extracting skeletons. This paper presents an effcient 3D parallel thinning algorithm which produces medial surfaces. A 3-subiteration strategy is proposed: the thinning operation is changed from iteration to iteration with a period of 3 according to the three deletion directions. [bibtex-key = PalagyiPRL2002]


64. Erich Sorantin, Csongor Halmai, Balazs Erdohelyi, Kalman Palagyi, Laszlo G. Nyul, Krisztian Olle, Bernhard Geiger, Franz Lindbichler, Gerhard Friedrich, and Karl Kiesler. Spiral-CT-based Assessment of Tracheal Stenoses Using 3-D-skeletonization. IEEE Transactions on Medical Imaging, 21(3):263-273, March 2002. [PDF]
    Abstract: Purpose: Demonstration of a technique for three-dimensional (3-D) assessment of tracheal-stenoses, regarding site, length and degree, based on spiral computed tomography (S-CT). Patients and Methods: S-CT scanning and automated segmentation of the laryngo-tracheal tract (LTT) was followed by the extraction of the LTT medial axis using a skeletonization algorithm. Orthogonal to the medial axis the LTT 3-D cross-sectional profile was computed and presented as line charts, where degree and length was obtained. Values for both parameters were compared between 36 patients and 18 normal controls separately. Accuracy and precision was derived from 17 phantom studies. Results: Average degree and length of tracheal stenoses was found to be 60.5% and 4.32 cm in patients compared with minor caliber changes of 8.8% and 2.31 cm in normal controls ( p<0.0001). For the phantoms an excellent correlation between the true and computed 3-D cross-sectional profile was found ( p < 0.005) and an accuracy for length and degree measurements of 2.14 mm and 2.53% respectively could be determined. The corresponding figures for the precision were found to be 0.92 mm and 2.56%. Conclusion: LTT 3-D cross-sectional profiles permit objective, accurate and precise assessment of LTT caliber changes. Minor LTT caliber changes can be observed even in normals and, in case of an otherwise normal S-CT study, can be regarded as artifacts. [bibtex-key = Sorantin:2002:SAT]


65. Erich Sorantin, Georg Werkgartner, Emese Balogh, Anna Vilanova i Bartroli, Kalman Palagyi, Laszlo G. Nyul, and Laszlo Rusko. Virtual dissection and automated polyp detection of the colon based on spiral CT data --- Techniques and preliminary experience on a cadaveric phantom. European Surgery -- Acta Chirurgica Austriaca, 34:143-149, 2002. [WWW] [doi:10.1046/j.1563-2563.2002.02018.x]
    Abstract: Background: CT colonography was found to be sensitive and specific for detection of colonic polyps and colorectal cancer (CRC). Depending on the software used, CT colonography requires a certain amount of operator interaction, which limits it's widespread usage. The goal of this papers is to present two novel automated techniques for displaying CT colonography: virtual dissection and automated colonic polyp detection. Methods: Virtual dissection refers to a technique where the entire colon is virtually stretched and flattened thus simulating the view on the pathologist's table. Colonic folds show a ``global outward bulging of the contour'', whereas colonic polyps exhibit the inverse (``local inward bulging''). This feature is used to map areas of ``local inward bulging'' with colours on 3D reconstructions. A cadaveric phantom with 13 artificially inserted polyps was used for validation of both techniques. Results: On virtual dissection all 13 inserted polyps could be identified. They appeared either as bumps or as local broadening of colonic folds. In addition, the automated colonic polyp detection algorithm was able to tag all polyps. Only 10 min of operator interaction were necessary for both techniques. Conclusions: Virtual dissection overcomes the shortcomings of CT colonography, and automated colonic polyp detection establishes a roadmap of the polyps. [bibtex-key = SorantinEtalACA2002]


66. Josiane Zerubia, Andre Jalobeanu, and Zoltan Kato. Markov random fields in image processing, application to remote sensing and astrophysics. Journal de Physique, EDP Sciences, IV(12):117-136, 2002. [WWW] [bibtex-key = Zerubia-etal2002]


67. Emese Balogh, Attila Kuba, Csaba Devenyi, and Alberto Del Lungo. Comparison of algorithms for reconstructing hv-convex discrete sets. Linear Algebra and its Applications, 339:23-35, December 2001. [PDF] [doi:10.1016/S0024-3795(01)00430-X]
    Abstract: Three reconstruction algorithms to be used for reconstructing hv-convex discrete sets from their row and column sums are compared. All these algorithms have two versions: one for reconstructing hv-convex polyominoes and another one for reconstructing hv-convex 8-connected discrete sets. In both classes of discrete sets the algorithms are compared from the viewpoints of average execution time and memory complexities. Discrete sets with given sizes are generated with uniform distribution, and then reconstructed from their row and column sums. First we have implemented two previously published algorithms. According to our comparisons, the algorithm which was better from the viewpoint of worst time complexity was the worse from the viewpoint of average time complexity and memory requirements. Then, as a new method, a combination of the two algorithms was also implemented and it is shown that it inherits the best properties of the other two methods. [bibtex-key = Balogh2001]


68. Sara Brunetti, Alberto DelLungo, F. DelRistoro, Attila Kuba, and Maurice Nivat. Reconstruction of 4- and 8-connected convex discrete sets from row and column projections. Linear Algebra and its Applications, 339:37-57, 2001. [PDF] [doi:10.1016/S0024-3795(01)00435-9]
    Abstract: In this paper we examine the problem of reconstructing a discrete two-dimensional set from its two orthogonal projection (H,V) when the set satisfies some convexity conditions. We show that the algorithm of the paper [Int. J. Imaging Systems and Technol. 9 (1998) 69] is a good heuristic algorithm but it does not solve the problem for all (H,V) instances. We propose a modification of this algorithm solving the problem for all (H,V) instances, by starting to build the ``spine''. The complexity of our reconstruction algorithm is O(mnˇlog(mn)ˇmin{m2,n2}) in the worst case. However, according to our experimental results, in 99% of the studied cases the algorithm is able to reconstruct a solution without using the newly introduced operation. In such cases the upper bound of the complexity of the algorithm is O(mnˇlog(mn)). A systematic comparison of this algorithm was done and the results show that this algorithm has the better average complexity than other published algorithms. The way of comparison and the results are given in a separate paper [Linear Algebra Appl. (submitted)]. Finally we prove that the problem can be solved in polynomial time also in a class of discrete sets which is larger than the class of convex polyominoes, namely, in the class of 8-connected convex sets. [bibtex-key = Brunetti2001]


69. Yulin Ge, Robert I. Grossman, Jayaram K. Udupa, James S. Babb, Laszlo G. Nyul, and Dennis L. Kolson. Brain Atrophy in Relapsing-remitting Multiple Sclerosis: Fractional Volumetric Analysis of Gray Matter and White Matter. Radiology, 220(3):606-610, September 2001. [PDF]
    Abstract: PURPOSE: To determine the fractional brain tissue volume changes in the gray matter and white matter of patients with relapsing-remitting multiple sclerosis (MS) and to correlate these measurements with clinical disability and total lesion load. MATERIALS AND METHODS: Thirty patients with relapsing-remitting MS and 25 healthy control subjects underwent magnetic resonance imaging. Fractional brain tissue volumes (tissue volume relative to total intracranial volume) were obtained from the total segmented gray matter and white matter in each group and were analyzed. RESULTS: The fractional volume of white matter versus that of gray matter was significantly lower (26.4%) in patients with MS (P , .0001) than in control subjects. Neither gray matter nor white matter fractional volume measurements correlated with clinical disability in the patients with MS. CONCLUSION: Loss of brain parenchymal volume in patients with relapsing-remitting MS is predominantly confined to white matter. Analysis of fractional brain tissue volumes provides additional information useful in characterizing MS and may have potential in evaluating treatment strategies. [bibtex-key = Ge:2001:BAR]


70. Zoltan Kato, Ting Chuen Pong, and John Chung Mong Lee. Color Image Segmentation and Parameter Estimation in a Markovian Framework. Pattern Recognition Letters, 22(3-4):309-321, March 2001. [PDF] Keyword(s): Unsupervised image segmentation, Color, Markov random fields, Pixel classification, Parameter estimation.
    Abstract: An unsupervised color image segmentation algorithm is presented, using a Markov random field (MRF) pixel classification model. We propose a new method to estimate initial mean vectors effectively even if the histogram does not have clearly distinguishable peaks. The only parameter supplied by the user is the number of classes. [bibtex-key = Kato-etal2001]


71. Attila Kuba and Maurice Nivat. Reconstruction of discrete sets with absorption. Linear Algebra and its Applications, 339:171-194, 2001. [WWW] [PDF] [doi:10.1016/S0024-3795(01)00486-4]
    Abstract: The uniqueness problem is considered when binary matrices are to be reconstructed from their absorbed row and column sums. Let the absorption coefficient n be selected such that en = (1+5^0.5)/2. Then it is proved that if a binary matrix is non-uniquely determined, then it contains a special pattern of 0s and 1s called composition of alternatively corner-connected components. In a previous paper [Discrete Appl. Math. (submitted)] we proved that this condition is also sufficient, i.e., the existence of such a pattern in the binary matrix is necessary and sufficient for its non-uniqueness. [bibtex-key = Kuba2001a]


72. Kalman Palagyi. A 3D parallel shrinking algorithm. Acta Cybernetica, 15:201-211, 2001. [PDF]
    Abstract: Shrinking is a frequently used preprocessing step in image processing. This paper presents an efficient 3D parallel shrinking algorithm for transforming a binary object into its topological kernel. The applied strategy is called directional: each iteration step is composed of six subiterations each of which can be executed in parallel. The algorithm makes easy implementation possible, since deletable points are given by 3x3x3 matching templates. The topological correctness of the algorithm is proved for (26,6) binary pictures. [bibtex-key = PalagyiAC2001]


73. Attila Tanacs, Gabor Czedli, Kalman Palagyi, and Attila Kuba. Affine matching of two sets of points in arbitrary dimensions. Acta Cybernetica, 15:101-106, 2001. [PDF]
    Abstract: In many applications of computer vision, image processing, and remotely sensed data processing, an appropriate matching of two sets of points is required. Our approach assumes one--to--one correspondence between these sets and finds the optimal global affine transformation that matches them. The suggested method can be used in arbitrary dimensions. A sufficient existence condition for a unique transformation is given and proven. [bibtex-key = Tanacs:2001:Acta]


74. Jayaram K. Udupa, Laszlo G. Nyul, Yulin Ge, and Robert I. Grossman. Multiprotocol MR Image Segmentation in Multiple Sclerosis: Experience With Over 1000 Studies. Academic Radiology, 8(11):1116-1126, November 2001. [PDF]
    Abstract: Rationale and Objectives. Multiple sclerosis (MS) is an acquired disease of the central nervous system. Several clinical measures are commonly used to express the severity of the disease, including the Expanded Disability Status Scale and the ambulation index. These measures are subjective and may be difficult to reproduce. The aim of this research is to investigate the possibility of developing more objective measures derived from MR imaging. Materials and Methods. Various magnetic resonance (MR) imaging protocols are being investigated for the study of MS. Seeking to replace the Expanded Disability Status Scale and ambulation index with an objective means to assess the natural course of the disease and its response to therapy, the authors have developed multiprotocol MR image segmentation methods based on fuzzy connectedness to quantify both macrosopic features of the disease (lesions, gray matter, white matter, cerebrospinal fluid, and brain parenchyma) and the microscopic appearance of diseased white matter. Over 1,000 studies have been processed to date. Results. By far the strongest correlations with the clinical measures were demonstrated by the magnetization transfer ratio histogram parameters obtained for the various segmented tissue regions. These findings emphasize the importance of considering the microscopic and diffuse nature of the disease in the individual tissue regions. Brain parenchymal volume also demonstrated a strong correlation with clinical measures, which suggests that brain atrophy is an important disease indicator. Conclusion. Fuzzy connectedness is a viable, highly reproducible segmentation method for studying MS. [bibtex-key = Udupa:2001:MMI]


75. Isabelle Catalaa, Robert I. Grossman, Dennis L. Kolson, Jayaram K. Udupa, Laszlo G. Nyul, Luogang Wei, Xuan Zhang, Marcia Polansky, Lois J. Mannon, and Joseph C. McGowan. Multiple Sclerosis: Magnetization Transfer Histogram Analysis of Segmented Normal-appearing White Matter. Radiology, 216(2):351-355, August 2000. [PDF]
    Abstract: PURPOSE: To investigate and characterize the global distribution of magnetization transfer (MT) ratio values of normal-appearing white matter (NAWM) in patients with relapsing-remitting multiple sclerosis (MS) and test the hypothesis that the MT histogram for NAWM reflects disease progression. MATERIALS AND METHODS: Conventional and MT magnetic resonance (MR) images were obtained in 23 patients and 25 healthy volunteers. Clinical tests for comparison with the MT histogram parameters included the Extended Disability Status Scale and the ambulation index. Lesion load calculated with T2-weighted MR images and whole-brain and white matter volumes were measured. RESULTS: The location of the MT histogram peak and the mean MT ratio for NAWM were significantly lower in patients with MS than in control subjects. In longitudinal studies, the histogram peak location and mean MT ratio shifted in the direction of normal values as the duration of disease increased. A mean of 26.5% of the volume of new lesions identified on the later studies were demonstrated to have originated in NAWM corresponding to ``lost'' pixels on the histogram. CONCLUSION: MT histogram analysis of NAWM, including longitudinal analysis, may provide new prognostic information regarding lesion formation and increase understanding of the course of the disease. [bibtex-key = Catalaa:2000:MSM]


76. Yulin Ge, Jayaram K. Udupa, Laszlo G. Nyul, Luogang Wei, and Robert I. Grossman. Numerical Tissue Characterization in MS via Standardization of the MR Image Intensity Scale. Journal of Magnetic Resonance Imaging, 12(5):715-721, November 2000. [PDF]
    Abstract: Image intensity standardization is a recently developed postprocessing method that is capable of correcting the signal intensity variations in MR images. We evaluated signal intensity of healthy and diseased tissues in 10 multiple sclerosis (MS) patients based on standardized dual fast spin-echo MR images using a numerical postprocessing technique. The main idea of this technique is to deform the volume image histogram of each study to match a standard histogram and to utilize the resulting transformation to map the image intensities into standard scale. Upon standardization, the coefficients of variation of signal intensities for each segmented tissue (gray matter, white matter, lesion plaques, and diffuse abnormal white matter) in all patients were significantly smaller (2.3-9.2 times) than in the original images, and the same tissues from different patients looked alike, with similar intensity characteristics. Numerical tissue characterizability of different tissues in MS achieved by standardization offers a fixed tissue-specific meaning for the numerical values and can significantly facilitate image segmentation and analysis. [bibtex-key = Ge:2000:NTC]


77. Laszlo G. Nyul and Jayaram K. Udupa. MR Image Analysis in Multiple Sclerosis. Advances in Multiple Sclerosis, 10(4):799-816, November 2000. [bibtex-key = Nyul:2000:MIA]


78. Laszlo G. Nyul, Jayaram K. Udupa, and Xuan Zhang. New Variants of a Method of MRI Scale Standardization. IEEE Transactions on Medical Imaging, 19(2):143-150, February 2000. [PDF]
    Abstract: One of the major drawbacks of magnetic resonance imaging (MRI) has been the lack of a standard and quantifiable interpretation of image intensities. Unlike in other modalities, such as X-ray computerized tomography, MR images taken for the same patient on the same scanner at different times may appear different from each other due to a variety of scanner-dependent variations and, therefore, the absolute intensity values do not have a fixed meaning.We have devised a two-step method wherein all images (independent of patients and the specific brand of the MR scanner used) can be transformed in such a way that for the same protocol and body region, in the transformed images similar intensities will have similar tissue meaning. Standardized images can be displayed with fixed windows without the need of per-case adjustment. More importantly, extraction of quantitative information about healthy organs or about abnormalities can be considerably simplified. This paper introduces and compares new variants of this standardizing method that can help to overcome some of the problems with the original method. [bibtex-key = Nyul:2000:NVM]


79. Tamas Sziranyi, Josiane Zerubia, Laszlo Czuni, David Geldreich, and Zoltan Kato. Image Segmentation Using Markov Random Field Model in Fully Parallel Cellular Network Architectures. Real Time Imaging, 6(3):195-211, June 2000. [PDF]
    Abstract: Markovian approaches to early vision processes need a huge amount of computing power. These algorithms can usually be implemented on parallel computing structures. Herein, we show that the Markovian labeling approach can be implemented in fully parallel cellular network architectures, using simple functions and data representations. This makes possible to implement our model in parallel imaging VLSI chips. As an example, we have developed a simplified statistical image segmentation algorithm for the Cellular Neural/Nonlinear Networks Universal Machine (CNN-UM), which is a new image processing tool, containing thousands of cells with analog dynamics, local memories and processing units. The Modified Metropolis Dynamics (MMD) optimization method can be implemented into the raw analog architecture of the CNN-UM. We can introduce the whole pseudo-stochastic segmentation process in the CNN architecture using 8 memories/cell. We use simple arithmetic functions (addition, multiplication), equality-test between neighboring pixels and very simple nonlinear output functions (step, jigsaw). With this architecture, the proposed VLSI CNN chip can execute a pseudo-stochastic relaxation algorithm of about 100 iterations in about 100 s. In the suggested solution the segmentation is unsupervised, where a pixel-level statistical estimation model is used. We have tested different monogrid and multigrid architectures. In our CNN-UM model several complex preprocessing steps can be involved, such as texture-classification or anisotropic diffusion. With these preprocessing steps, our fully parallel cellular system may work as a high-level image segmentation machine, using only simple functions based on the close-neighborhood of a pixel. [bibtex-key = Sziranyi-etal2000]


80. Attila Kuba. Reconstruction of two-valued functions and matrices. In Gabor T. Herman and Attila Kuba, editors, Discrete Tomography: Foundations, Algorithms, and Applications, Applied and Numerical Harmonic Analysis, pages 133-158. Birkhauser, Boston, 1999. [bibtex-key = Kubabirk1999]


81. Attila Kuba and Gabor T. Herman. Discrete Tomography: A historical overview. In Gabor T. Herman and Attila Kuba, editors, Discrete Tomography: Foundations, Algorithms, and Applications, Applied and Numerical Harmonic Analysis, chapter 1, pages 1-30. Birkhauser, Boston, 1999. [bibtex-key = KubaHerman19991]


82. Zoltan Kato, Josiane Zerubia, and Mark Berthod. Unsupervised Parallel Image Classification Using Markovian Models. Pattern Recognition, 32(4):591-604, April 1999. [PDF] Keyword(s): Markov random field model, Hierarchical model, Parameter estimation, Parallel unsupervised image classification.
    Abstract: This paper deals with the problem of unsupervised classification of images modeled by Markov random fields (MRF). If the model parameters are known then we have various methods to solve the segmentation problem (simulated annealing (SA), iterated conditional modes (ICM), etc). However, when the parameters are unknown, the problem becomes more difficult. One has to estimate the hidden label field parameters only from the observed image. Herein, we are interested in parameter estimation methods related to monogrid and hierarchical MRF models. The basic idea is similar to the expectation-maximization (EM) algorithm: we recursively look at the maximum a posteriori (MAP) estimate of the label field given the estimated parameters, then we look at the maximum likelihood (ML) estimate of the parameters given a tentative labeling obtained at the previous step. The only parameter supposed to be known is the number of classes, all the other parameters are estimated. The proposed algorithms have been implemented on a Connection Machine CM200. Comparative experiments have been performed on both noisy synthetic data and real images. [bibtex-key = Kato-etal99]


83. Laszlo G. Nyul and Jayaram K. Udupa. On Standardizing the MR Image Intensity Scale. Magnetic Resonance in Medicine, 42(6):1072-1081, December 1999. [PDF]
    Abstract: The lack of a standard image intensity scale in MRI causes many difficulties in image display and analysis. A two-step postprocessing method is proposed for standardizing the intensity scale in such a way that for the same MR protocol and body region, similar intensities will have similar tissue meaning. In the first step, the parameters of the standardizing transformation are ``learned'' from a set of images. In the second step, for each MR study these parameters are used to map their histogram into the standardized histogram. The method was tested quantitatively on 90 whole-brain studies of multiple sclerosis patients for several protocols and qualitatively for several other protocols and body regions. Measurements using mean squared difference showed that the standardized image intensities have statistically significantly (P < 0.01) more consistent range and meaning than the originals. Fixed gray level windows can be established for the standardized images and used for display without the need of per case adjustment. Preliminary results also indicate that the method facilitates improving the degree of automation of image segmentation. [bibtex-key = Nyul:1999:OSM]


84. Kalman Palagyi and Attila Kuba. A parallel 3D 12-subiteration thinning algorithm. Graphical Models and Image Processing, 61:199-221, 1999. [PDF]
    Abstract: Thinning on binary images is an iterative layer by layer erosion until only the "skeletons" of the objects are left. This paper presents an efficient parallel thinning algorithm which produces either curve skeletons or surface skeletons from 3D binary objects. It is important that a curve skeleton is extracted directly (i.e., without creating a surface skeleton). The strategy which is used is called directional: each iteration step is composed of a number of subiterations each of which can be executed in parallel. One iteration step of the proposed algorithm contains 12 subiterations instead of the usual six. The algorithm makes easy implementation possible, since deletable points are given by 3L3L3 matching templates. The topological correctness for (26,6) binary pictures is proved. [bibtex-key = PalagyiKubaGMIP1999]


85. J. H. B. Kemperman and Attila Kuba. Reconstruction of two-valued matrices from their two projections. International Journal of Imaging Systems and Technology, 9(2-3):110-117, December 1998. [doi:10.1002/(SICI)1098-1098(1998)9:2/3<110::AID-IMA7>3.0.CO;2-E]
    Abstract: A matrix is said to be two-valued if its elements assume at most two different values. We studied the problem of reconstructing a two-valued matrix from its marginals - that is, from its row sums and column sums - but without any knowledge of the value pair on hand. Provided at least one of these marginals is nonconstant, only finitely many (though possibly many) value pairs can lead to a valid reconstruction. Our considerations lead to an efficient algorithm for calculating all possible solutions, each with its own value pair. Special attention is given to uniqueness pairs - that is, value pairs to which there corresponds precisely one matrix having the correct marginals. Unless both marginals are constant, there can be no more than two uniqueness pairs. [bibtex-key = Kemperman1998]


86. Juha Kivijarvi, Tiina Ojala, Timo Kaukoranta, Attila Kuba, Laszlo G. Nyul, and Olli Nevalainen. A comparison of lossless compression methods for medical images. Computerized Medical Imaging and Graphics, 22:323-339, 1998. [PDF] [doi:10.1016/S0895-6111(98)00042-1]
    Abstract: In this work, lossless grayscale image compression methods are compared on a medical image database. The database contains 10 different types of images with bit rates varying from 8 to 16 bits per pixel. The total number of test images was about 3000, originating from 125 different patient studies. Methods used for compressing the images include seven methods designed for grayscale images and 18 ordinary general-purpose compression programs. Furthermore, four compressed image file formats were used. The results show that the compression ratios strongly depend on the type of the image. The best methods turned out to be TMW, CALIC and JPEG-LS. The analysis step in TMW is very time-consuming. CALIC gives high compression ratios in a reasonable time, whereas JPEG-LS is nearly as effective and very fast. [bibtex-key = Kivijarvi1998]


87. Kalman Palagyi and Attila Kuba. A 3D 6-subiteration thinning algorithm for extracting medial lines. Pattern Recognition Letters, 19:613-627, 1998. [PDF]
    Abstract: Thinning is a frequently used method for extracting skeletons in discrete spaces. This paper presents an efficient parallel thinning algorithm that directly extracts medial lines from elongated 3D binary objects (i.e., without creating medial surface). Our algorithm provides good results, preserves topology and it is easy to implement. [bibtex-key = PalagyiKubaPRL1998]


88. Kalman Palagyi and Attila Kuba. A hybrid thinning algorithm for 3D medical images. Journal of Computing and Information Technology, 6:149-164, 1998. [PDF]
    Abstract: Thinning is a frequently used method for extracting skeletons in discrete spaces. This paper presents an efficient parallel algorithm for thinning elongated 3D binary objects (e.g., bony structures, vessel trees, or airway trees). The proposed algorithm directly extracts medial lines as shape features from 3D binary objects by applying a brand-new class of thinning strategy called hybrid method. Our topology preserving algorithm makes easy implementation possible and gives satisfactory results for synthetic data tests and for MR angiography brain studies. [bibtex-key = PalagyiKubaCIT1998]


89. R. Aharoni, Gabor T. Herman, and Attila Kuba. Binary vectors partially determined by linear equation systems. Discrete Mathematics, 171(1-3):1-16, June 1997. [PDF] [doi:10.1016/S0012-365X(96)00068-4]
    Abstract: We address problems of the general form: given a J-dimensional binary (0- or 1-valued) vector a, a system of E of linear equations which a satisfies and a domain D subset of RJ which contains a, when is a the unique solution of E in D? More generally, we aim at finding conditions for the invariance of a particular position j, 1 less-than-or-equals, slant j less-than-or-equals, slant J (meaning that bj = aj, for all solutions b of E in D. We investigate two particular choices for D: the set of binary vectors of length J (integral invariance) and the set of vectors in RJ whose components lie between 0 and 1 (fractional invariance). For each position j, a system of inequalities is produced, whose solvability in the appropriate space indicates variance of the position. A version of Farkas' Lemma is used to specify the alternative system of inequalities, giving rise to a vector using which one can tell for each position whether or not it is fractionally invariant. We show that if the matrix of E is totally unimodular, then integral invariance is equivalent to fractional invariance. Our findings are applied to the problem of reconstruction of two-dimensional binary pictures from their projections (equivalently, (0, 1)-matrices from their marginals) and lead to a “structure result” on the arrangement of the invariant positions in the set of all binary pictures which share the same row and column sums and whose values are possibly prescribed at some positions. The relationship of our approach to the problem of reconstruction of higher-dimensional binary pictures is also discussed. [bibtex-key = Aharoni1997]


90. Mark Berthod, Zoltan Kato, Shan Yu, and Josiane Zerubia. Bayesian Image Classification Using Markov Random Fields. Image and Vision Computing, 14:285-295, 1996. [PDF] Keyword(s): Bayesian image classification, Markov random fields, Optimisation.
    Abstract: In this paper, we present three optimisation techniques, Deterministic Pseudo-Annealing (DPA), Game Strategy Approach (GSA), and Modified Metropolis Dynamics (MMD), in order to carry out image classification using a Markov random field model. For the first approach (DPA), the a posteriori probability of a tentative labelling is generalised to a continuous labelling. The merit function thus defined has the same maxima under constraints yielding probability vectors. Changing these constraints convexities the merit function. The algorithm solves this unambiguous maximisation problem, and then tracks down the solution while the original constraints are restored yielding a good, even if suboptimal, solution to the original labelling assignment problem. In the second method (GSA), the maximisation problem of the a posteriori probability of the labelling is solved by an optimisation algorithm based on game theory. A non-cooperative n-person game with pure strategies is designed such that the set of Nash equilibrium points of the game is identical to the set of local maxima of the a posteriori probability of the labelling. The algorithm converges to a Nash equilibrium. The third method (MMD) is a modified version of the Metropolis algorithm: at each iteration the new state is chosen randomly, but the decision to accept it is purely deterministic. This is also a suboptimal technique but it is much faster than stochastic relaxation. These three methods have been implemented on a Connection Machine CM2. Experimental results are compared to those obtained by the Metropolis algorithm, the Gibbs sampler and ICM (Iterated Conditional Mode). [bibtex-key = Berthod-etal96]


91. Zoltan Kato, Mark Berthod, and Josiane Zerubia. A Hierarchical Markov Random Field Model and Multi-Temperature Annealing for Parallel Image Classification. Computer Vision, Graphics and Image Processing: Graphical Models and Image Processing, 58(1):18-37, January 1996. [PDF] [PS]
    Abstract: In this paper, we are interested in massively parallel multiscale relaxation algorithms applied to image classification. It is well known that multigrid methods can improve significantly the convergence rate and the quality of the final results of iterative relaxation techniques. First, we present a classical multiscale model which consists of a label pyramid and a whole observation field. The potential functions of coarser grids are derived by simple computations. The optimization problem is first solved at the higher scale by a parallel relaxation algorithm, then the next lower scale is initialized by a projection of the result. Second, we propose a hierarchical Markov Random Field model based on this classical model. We introduce new interactions between neighbor levels in the pyramid. It can also be seen as a way to incorporate cliques with far apart sites for a reasonable price. This model results in a relaxation algorithm with a new annealing scheme: The Multi-Temperature Annealing (MTA) scheme, which consists of associating higher temperatures to higher levels, in order to be less sensitive to local minima at coarser grids. The convergence to the global optimum is proved by a generalisation of the annealing theorem of Geman and Geman. [bibtex-key = Kato-etal96]


92. Mark Berthod, Zoltan Kato, and Josiane Zerubia. DPA: A Deterministic Approach to the MAP. IEEE Transactions on Image Processing, 4(9):1312-1314, September 1995.
    Abstract: Deterministic pseudo-annealing (DPA) is a new deterministic optimization method for finding the maximum a posteriori (MAP) labeling in a Markov random field, in which the probability of a tentative labeling is extended to a merit function on continuous labelings. This function is made convex by changing its definition domain. This unambiguous maximization problem is solved, and the solution is followed down to the original domain, yielding a good, if suboptimal, solution to the original labeling assignment problem. The performance of DPA is analyzed on randomly weighted graphs [bibtex-key = Berthod-etal95c]


93. Attila Kuba. Reconstruction of Unique Binary Matrices with Prescribed Elements. Acta Cybernetica, 12(1):57-70, 1995. [WWW]
    Abstract: The reconstruction of a binary matrix from its row and column sum vectors is considered when some elements of the matrix may be prescribed and the matrix is uniquely determined from these data. It is shown that the uniqueness of such a matrix is equivalent to the impossibility of selecting certain sequences from the matrix elements. The unique matrices are characterized by several properties. Among others it is proved that their rows and columns can be permutated such that the 1's are above and left to the (non-prescribed) 0's. Furthermore, an algorithm is given to decide if the given projections and prescribed elements determine a binary matrix uniquely, and, if the answer is yes, to reconstruct it. [bibtex-key = Kuba375983]


94. Zoltan Kato, Josiane Zerubia, and Mark Berthod. Bayesian Image Classification Using Markov Random Fields. In Ali Mohammad-Djafari and Guy Demoment, editors, Maximum Entropy and Bayesian Methods, pages 375-382. Kluwer Academic Publisher, 1993.
    Abstract: In this paper, we present two relaxation techniques: Deterministic Pseudo-Annealing (DPA) and Modified Metropolis Dynamics (MMD) in order to do image classification using a Markov Random Field modelization. For the first algorithm (DPA), the a posteriori probability of a tentative labeling is generalized to continuous labeling. The merit function thus defined has the same maxima under constraints yielding probability vectors. Changing these constraints convexify the merit function. The algorithm solve this unambigous maximization problem and then tracks down the solution while the original constraints are restored yielding a good even if suboptimal solution to the original labeling assignment problem. As for the second method (MMD), it is a modified version of the Metropolis algorithm: at each iteration the new state is chosen randomly but the decision to accept it is purely deterministic. This is of course also a suboptimal technique which gives faster results than stochastic relaxation. These two methods have been implemented on a Connection Machine CM2 and simulation results are shown with a SPOT image and two synthetic noisy images. These results are compared to those obtained with the Metropolis algorithm, the Gibbs sampler and ICM (Iterated Conditional Mode). [bibtex-key = Kato-etal93]


95. Attila Kuba and A. Volcic. The structure of the class of non-uniquely reconstructible sets. Acta Scientiarum Mathematicarum, 58(1-4):359-384, 1993. [WWW]
    Abstract: Consider the class of all measurable plane sets having given horizontal and vertical projections. According to this class the plane can be divided into three subsets: the essentially common subset of all elements of the class, the essentially common subset of the complements of all elements of the class, and the remaining subset of the plane. The three sets together are called the structure of the class. In this paper a method is given by which the structure of an arbitrary class can be determined from the projections. The method is similar to the procedure applied in the case of binary matrices. First, the structure of the normalized class (having rearranged non-increasing projections from the original ones) is constructed. Then, by a measure preserving mapping the structure of the original class is derived from the structure of the normalized class. The structure can be used in the reconstruction of non-unique sets from their projections, e.g. it gives information about the shape and the position of the possible solutions and an upper bound of the measure of the difference between two solutions. [bibtex-key = KubaVolcic199358]


96. Endre Katona. Cellular Processing. In Branco Soucek and the IRIS Group, editors, Fuzzy, Holographic and Parallel Intelligence, pages 215-230. John Wiley & Sons, Inc., New York, USA, 1992. [PDF]
    Abstract: How can we use a homogeneous array of elementary processors (cells) to solve different computational problems? An early answer to this question was the cellular automata concept of J. von Neumann. Later on cellular arrays appeared in (V)LSI circuit design, in Programmable Gate Arrays, different cellular automata realizations have been proposed, and the notion of systolic arrays was created. On the other hand, different massively parallel machines have been constructed (MPP, DAP, GAPP, etc.) based on cellular computing too. This chapter gives an overview of cellular processing with special respect to cellular processors developed by Cellware Ltd. Section 6.2 sketches a general cellular array model forming a common theoretical frame for very different cellular array machines. Sections 6.3 and 6.4 describe basic hardware implementation techniques of cellular arrays illustrated on cellprocessors M1, X1 and SP1 of Cellware Ltd. Sections 6.5 and 6.6 discuss how to realize pipeline processing in cellular arrays. Problems of modeling, partitioning and I/O management are examined. The different cellular array machines are compared on the basis of the general model of Section 6.2, connections are discovered between inhomogeneity, efficiency and programming methodology. [bibtex-key = Katona1992]


97. Attila Kuba, Arpad Makay, Eors Mate, and Laszlo Csernay. Data processing system for nuclear medicine images. International Journal of Imaging Systems and Technology, 4:51-61, 1992. [PDF] [doi:10.1002/ima.1850040111]
    Abstract: A software system (MicroSegams) is described which is able to acquire, store, present, and process nuclear medicine images. The most important basic procedures (e.g., image acquisition from a gamma camera, selection of a study from the catalogue, drawing regions of interest, etc.) and complete clinical procedures (e.g., renography, brain perfusion, etc.) can be activated directly from the system menu by mouse. The language sLan (a subset of programming language C containing more than 200 elementary functions) can also be used to write new procedures to be interpreted or compiled. The system is offered as a basic tool for the performance of static, dynamic, gated, or tomographic investigations in clinical practice. [bibtex-key = Kuba1992]


98. Mark F. Lythgoe, Hugh Davies, Attila Kuba, Mihaly Toth-Abonyi, and Isky Gordon. Can Dynamic Krypton-81m Imaging Separate Regional Ventilation and Volume?. The Journal of Nuclear Medicine, 33(11):1935-1939, 1992. [PDF]
    Abstract: This study explores the assumption that 81mKr static images represent regional ventilation. Dynamic acquisition of 81mKr ventilation images permits creation of time-activity curves and the possible separation of the confounding influences of ventilation and volume. By using a two-compartment gas mixing lung phantom, the results demonstrate that both total and tidal 81mKr are closely related to regional ventilation. In 61 children and 15 adult volunteers, there was good agreement between fractional ventilation assessed by total and tidal 81mKr. The dynamic steady-state ventilation image can be analyzed to separate tidally exchanged and resident 81mKr. This may allow regional ventilation to be distinguished from regional volume. [bibtex-key = Lythgoe1992]


99. Eors Mate, Janos Mester, Laszlo Csernay, Attila Kuba, Madani Shahram, and Arpad Makay. Three-dimensional presentation of the Fourier amplitude and phase: a fast display method for gated cardiac blood-pool SPECT. The Journal of Nuclear Medicine, 33(3):458-462, March 1992. [WWW] [PDF]
    Abstract: The routine clinical use of gated SPECT is inhibited by sophisticated, time-consuming processing techniques. The present paper describes a new technique for the simultaneous three-dimensional presentation of the amplitude and phase of the first Fourier harmonics, with the aim of obtaining detailed information about the ventricular motion in a relatively short time, from each angle of view of three-dimensional space. The method is simple and robust, and processing is automatic. It does not need carefully elaborated techniques for surface determination, because the cardiac surface is merely used as a reference skeleton onto which the functional information of amplitude and phase is mapped. The Fourier analysis before reconstruction results in running times shorter than 15 min and may further open the way for the routine use of gated SPECT. [bibtex-key = Mate1992]


100. Attila Kuba. Reconstruction of measurable plane sets from their two projections taken in arbitrary directions. Inverse Problems, 7(1):101-107, 1991. [PDF] [doi:10.1088/0266-5611/7/1/010]
     Abstract: The problem of the reconstruction of a measurable plane set from two projections is considered when the directions of the projections are arbitrary. Via a suitable affine transformation, the problem can be reduced to the solved case of orthogonal projections. The result can also be applied for unknown directions. It is proved that merely a knowledge of the projections (i.e. without their directions) is not enough for the unique reconstruction of a set. However, from two given functions it is possible to decide whether they can determine a set from some directions uniquely, and the corresponding directions of the projections are also computable. [bibtex-key = Kuba1991a]


101. Dietrich Kolzow, Attila Kuba, and A. Volcic. An algorithm for reconstructing convex bodies from their projections. Discrete and Computational Geometry, 4(1):205-237, December 1989. [doi:10.1007/BF02187723]
     Abstract: An algorithm is described for the approximative reconstruction of a plane convex body from its projections in a finite number of directions.A priori anda posteriori error estimates are given, and the convergence of certain sequences of an approximative solution of the reconstruction problem to the exact solution is proven. Finally, it is shown that, after small modifications, the algorithm can be applied to reconstruct convex bodies from discrete projectional data. The algorithm consists in an approximation of the convex body, to be reconstructed, by recursively defined cores and envelopes, following the ideas of Kuba [6] for the reconstruction of binary patterns. This paper was started at the University of Erlangen-Nurnberg in 1983-1984, when the second author was a fellow of the Alexander von Humboldt Foundation, while the last author was supported by the German Research Council (DFG, Contract Ko-506/8-1). [bibtex-key = KolzowKuba]


102. Attila Kuba. Determination of the structure of the class A(R,S) of (0,1)-matrices. Acta Cybernetica, 9(2):121-132, 1989. [WWW] [PDF] [bibtex-key = Kuba248942]


103. Attila Kuba and A. Volcic. Characterisation of measurable plane sets which are reconstructable from their two projections. Inverse Problems, 4(2):513-527, May 1988. [PDF] [doi:10.1088/0266-5611/4/2/014]
     Abstract: The authors complement previous work of Lorentz (1949) on the determination of plane measurable sets from two projections. They give three different characterisations of uniquely determined sets and show a reconstruction formula for such sets. [bibtex-key = Kuba1988h4]


104. Tamas Legendi, Endre Katona, Jozsef Toth, and Antal Zsoter. Megacell Machine. Parallel Computing, 8:195-199, 1988. [PDF]
     Abstract: A general purpose highly parallel cellular processor, called Megacell Machine, has been studied and developed. This paper presents the architecture of the Megacell Machine, discusses the possibilities of LSI, VLSI, WLSI implementation, sketches its programming methodology and programming languages. Some application fields (image processing, matrix operations, finite element analysis, etc.) and performance data will be mentioned. [bibtex-key = Legendi1988]


105. Janos Csirik and Eors Mate. The probabilistic behaviour of the NFD bin packing algorithm. Acta Cybernetica, 7:241-245, 1986. [PDF] [bibtex-key = Csirik1986]


106. Endre Katona. A lattice model for cellular (systolic) algorithms. Parallel Computing, 3:251-258, 1986. [PDF]
     Abstract: In this paper a lattice model (here ``lattice'' means a net of points) is introduced for homogeneous cellular algorithms. On the basis of this model a transformation methodology is developed, which makes it possible to produce many different versions of a given cellular algorithm. These versions may have quite different structural properties, but they perform the same computation as the original algorithm. In this way a great variety of cellular algorithms can be offered to choose the best version in practice and, on the other hand, cellular algorithms can be classified according to their inherent structures. [bibtex-key = Katona1986b]


107. Janos Mester, I Preda, G Kozmann, Eors Mate, L Regos, K Toth, Gyorgy Marosi, Zoltan Antaloczy, and Laszlo Csernay. Detection and localisation of ventricular preexcitation in Wolff-Parkinson-White syndrome (comparative study with body surface mapping). Nuclear Medicine Communications, 7:497-504, July 1986. [PDF]
     Abstract: 21 patients with Wolff-Parkinson-White (WPW) syndrome were investigated through radionuclide imaging and body surface mapping. Ventricular preexcitation was localized by display of identical phase ventricular regions (phase display). In 79% of the cases radioisotope and body surface mapping methods have identical results for the site of preexcitation. In two patients of the control group (10 patients with Lown-Ganong-Levine (LGL) syndrome), false positive outside-of septum ventricular preexcitation was detected, i.e. an 80% specificity. It is concluded that phase display might be a suitable method in the diagnosis of WPW syndrome. The sensitivity of this method can be improved using several views. [bibtex-key = Mester1986]


108. Jozsef Bodis, Katalin Zambo, Zoltan Nemessanyi, Eors Mate, and Imre F. Csaba. Application of the parametric scan in the investigation of uteroplacental blood flow. European Journal of Nuclear Medicine and Molecular Imaging, 10(5):286-287, March 1985. [PDF] [doi:10.1007/BF00254477]
     Abstract: We developed a new radioisotope technique to measure placental blood flow for the early detection of placental insufficiency. Using the parametric scan, placental perfusion was measured in 80 late-pregnant women. The T-maximum pictures obtained made it possible to differentiate between the vascular and intervillous phases of placental blood flow. The time period of intervillous phase calculated as the percentage of the whole placental T max was given as the intervillous perfusion index (IPI). It was demonstrated that the IPI was significantly higher in pregnancies complicated with intrauterine growth retardation (64.2%±16.5%), hypertension (60.0±15.2) and oedema (57.9%±10.1%) than in the control group (33.7%±10.5%). These data suggest that the first sign of placental insufficiency is the prolongation of the IPI which is likely to precede the quantitative reduction of placental perfusion. [bibtex-key = Bodis1985]


109. Jozsef Bodis, Katalin Zambo, Zoltan Nemessanyi, Eors Mate, and Imre F. Csaba. Determination of the intervillous perfusion index in pregnancies with intrauterine growth retardation. Gynecologic and Obstetric Investigation, 19:89-91, 1985. [PDF] [bibtex-key = Bodis1985a]


110. Attila Kuba. The reconstruction of two-directionally connected binary patterns from their two orthogonal projections. Computer Vision, Graphics, and Image Processing, 27(3):249-265, September 1984. [doi:10.1016/0734-189X(84)90031-8]
     Abstract: A heuristic reconstruction algorithm is described, by which binary matrices can be produced from their two orthogonal projections. It is necessary for proper reconstruction that the binary pattern corresponding to the binary matrix bex- andy-directionally connected. By this method a large class of binary matrices can be reconstructed. It is proved that after a finite number of iterative steps this algorithm produces all thex- andy-directionally connected binary patterns belonging to the given two projections. Finally, some remarks on the implementation of this algorithm and results are presented. [bibtex-key = Kuba1984]


111. Attila Kuba, H. Feistel, J. Mahlstedt, and F. Wolf. 3D ROI techniques in SPECT. European Journal of Nuclear Medicine and Molecular Imaging, 9:A130, 1984. [WWW] [bibtex-key = Kuba19849]


112. Tamas Legendi and Endre Katona. A Solution of the Early Bird Problem in an n-dimensional Cellular Space. Acta Cybernetica, 7:81-87, 1984. [PDF]
     Abstract: The early bird problem has been originally defined and solved by Rosenstiehl et al. (1972) and has also been solved by Vollmar (1978), Legendi and Katona (1981). In this paper the problem is generalized for the n-dimensional cellular space, and a real time solution is given using O(mn) steps in an n-dimensional cellular field of edge length m. The solution will be shown in detail in the two-dimensional case. The basic idea of the solution is to reduce the n-dimensional early bird problem to the (n-1)-dimensional one, using special signals and applying the one-dimensional early bird algorithm. [bibtex-key = Legendi1984]


113. Endre Katona. Bit-Parallel Processing with Bit-Sequential and Bit-Parallel Data Handling in Cellular Space. Computer and Artificial Intelligence, 2:315-329, 1983. [PDF]
     Abstract: In the first part of this paper the main principles of cellular processing are discussed, the notions of bit-sequential and bit-parallel data handling are introduced. Further, some fundamental cellular processing elements are presented for addition, multiplication and comparison with both bit-sequential and bit-parallel data input. To demonstrate the interconnection of the processing elements, more complex cellular algorithms are briefly shown as sorting, code translation, matrix multiplication. Finally, some conclusions are given concerning the bit-sequential and bit-parallel data handling. [bibtex-key = Katona1983]


114. Endre Katona. Binary Addition and Multiplication in Cellular Space. Acta Cybernetica, 5:457-464, 1982. [PDF]
     Abstract: Cellular automata are highly parallel bitprocessors, so they are suitable for the bitparallel execution of distinct computational tasks. In this paper powerful bitparallel algorithms are given for fixed point binary addition and multiplication, taking into account the cellprocessor architecture developed by T. Legendi. For this architecture there have been constructed more than 100 cellular algorithms solving different computational tasks. In a large cellular space a high number of cellular adders, multipliers and other processing elements may be embedded, and more complex tasks may be computed in parallel, as matrix multiplication, certain data processing tasks, etc. [bibtex-key = Katona1982]


115. Tamas Legendi and Endre Katona. A 5 State Solution of the Early Bird Problem in a One Dimensional Cellular Space. Acta Cybernetica, 5:173-179, 1981. [PDF]
     Abstract: There exists a class of interesting problems for cellular automata characterized by their common property of decomposing some global behaviour into homogeneous parallel transitions. Well known representatives of this class are the firing squad synchronization problem and the French Flag problem. Another problem of this class was defined by Rosenstiehl et al. (1972) named as the ``early bird'' problem. The problem has been solved also by Vollmar (1978) using a high number of cell states. The solution presented here uses only 5 cell states. [bibtex-key = Legendi1981]


116. Endre Katona. Linear Parallel Maps of Tesselation Automata. Acta Cybernetica, 4:263-268, 1979. [PDF]
     Abstract: The s-state cellular automaton, where the local map computes the modulo-s sum of all neighbour-states, has very useful algebraic properties. The principle of this cellular automaton will be generalized in this paper, as far as it is possible, applying inhomogeneous tessellation automata, so we get the concept of linear parallel maps. A wide class of parallel maps is obtained in such a way, keeping the good algebraic properties of the original automata. The present work discusses the fundamental characteristics of linear parallel maps, giving some examples and applications to demonstrate the theoretical results. Finally, we mention an open problem showing that many further investigations are possible in the area. [bibtex-key = Katona1979]


117. Attila Kuba and Laszlo Csernay. Simulatiuon studies in reconstruction tomography. European Journal of Nuclear Medicine and Molecular Imaging, 4(2):144, March 1979. [doi:10.1007/BF00626081] [bibtex-key = KubaCsernai1979jb]


118. Endre Katona. Local and Global Reversibility of Finite Inhomogeneous Cellular Automaton. Acta Cybernetica, 3:287-292, 1978. [PDF]
     Abstract: Cellular automata (CA) are highly parallel working systems, so they have high importance in computational applications. For this reason the classical definitions of finite, homogeneous CA are modified in this work: we introduce the notion of finite, inhomogeneous CA. The problem of reversibility has been investigated by many authors, but only for the global maps. To the contrary, we mean the reversibility in local sense: a CA we call reversible, if its local maps can be changed so that the new global map is the inverse of the original one. A connection is proved between the local maps and the number of Eden-configurations, from which derives a necessary condition for bijectivity of global maps. Next, a necessary and sufficient condition is presented to the reversibility. Finally, concrete investigations are given in the case of one-dimensional CA, with the result: only very simple reversible CA exist in this special case. [bibtex-key = Katona1978]


119. Jozsef Szucs, Eors Mate, and Ciprian Foias. Non-existence of unicellular operators on certain nonseparable Banach spaces. Acta Scientiarum Mathematicarum, 31:297-300, 1970. [WWW] [bibtex-key = Szucs1970]


120. Eors Mate. On a problem of P. Erdos. Acta Scientiarum Mathematicarum, 30:301-304, 1969. [WWW] [bibtex-key = Eoers1969]


121. Peter Kardos and Kalman Palagyi. On topology preservation for hexagonal parallel thinning algorithms. In J.K. Aggarwal, R.P. Barneva, V.E. Brimkov, K.N. Koroutchev, and E.R. Korutcheva, editors, Proceedings of the International Workshop on Combinatorial Image Analysis, volume 6636 of Lecture Notes in Computer Science, Madrid, Spain, pages 31-42, May 2011. Springer Verlag. [PDF]
     Abstract: Thinning is an iterative layer-by-layer erosion until only the skeleton-like shape features of the objects are left. This paper presents a family of new 3D parallel thinning algorithms that are based on our new sufficient conditions for 3D parallel reduction operators to preserve topology. The strategy which is used is called subiteration-based: each iteration step is composed of six parallel reduction operators according to the six main directions in 3D. The major contributions of this paper are: 1) Some new sufficient conditions for topology preserving parallel reductions are introduced. 2) A new 6-subiteration thinning scheme is proposed. Its topological correctness is guaranteed, since its deletion rules are derived from our sufficient conditions for topology preservation. 3) The proposed thinning scheme with different characterizations of endpoints yields various new algorithms for extracting centerlines and medial surfaces from 3D binary pictures. [bibtex-key = KP_IWCIA2011]


122. Peter Kardos and Kalman Palagyi. Order-independent sequential thinning in arbitrary dimensions. In Proceedings of the IASTED International Conference Signal and Image Processing and Applications, Crete, Greece, pages 129-134, June 2011. IASTED. [PDF] [doi:DOI: 10.2316/P.2011.738-021]
     Abstract: Skeletons are region based shape descriptors that play important role in shape representation. This paper introduces a novel sequential thinning approach for n-dimensional binary objects (n=1,2,3,...). Its main strength lies in its order-independency, i.e., it can produce the same skeletons for any visiting orders of border points. Furthermore, this is the first scheme in this field that is also applicable for higher dimensions. [bibtex-key = KP_SIPA2011]


123. Jhimli Mitra, Zoltan Kato, Robert Marti, Arnau Oliver, Xavier Llado, Soumya Ghose, Joan C. Vilanova, and Fabrice Meriaudeau. A Non-linear Diffeomorphic Framework for Prostate Multimodal Registration. In Proceedings of International Conference on Digital Image Computing: Techniques and Applications, Noosa, Queensland, Australia, pages 31-36, December 2011. IEEE. [PDF]
     Abstract: This paper presents a novel method for non-rigid registration of prostate multimodal images based on a nonlinear framework. The parametric estimation of the non-linear diffeomorphism between the 2D fixed and moving images has its basis in solving a set of non-linear equations of thin-plate splines. The regularized bending energy of the thin-plate splines along with the localization error of established correspondences is jointly minimized with the fixed and transformed image difference; where, the transformed image is represented by the set of non-linear equations defined over the moving image. The traditional thin-plate splines with established correspondences may provide good registration of the anatomical targets inside the prostate but may fail to provide improved contour registration. On the contrary, the proposed framework maintains the accuracy of registration in terms of overlap due to the non-linear thinplate spline functions while also producing smooth deformations of the anatomical structures inside the prostate as a result of established corrspondences. The registration accuracies of the proposed method are evaluated in 20 pairs of prostate midgland ultrasound and magnetic resonance images in terms of Dice similarity coefficient with an average of 0.982 ± 0.004, average 95% Hausdorff distance of 1.54 ± 0.46 mm and mean target registration and target localization errors of 1.90±1.27 mm and 0.15 ± 0.12 mm respectively. [bibtex-key = Mitra-etal2011]


124. Gabor Nemeth, Peter Kardos, and Kalman Palagyi. A family of topology-preserving 3D parallel 6-subiteration thinning algorithms. In J.K. Aggarwal, R.P. Barneva, V.E. Brimkov, K.N. Koroutchev, and E.R. Korutcheva, editors, Proceedings of the International Workshop on Combinatorial Image Analysis, volume 6636 of Lecture Notes in Computer Science, Madrid, Spain, pages 17-30, May 2011. Springer Verlag. [PDF]
     Abstract: Topology preservation is the key concept in parallel thinning algorithms on any sampling schemes. This paper establishes some sufficient conditions for parallel thinning algorithms working on hexagonal grids (or triangular lattices) to preserve topology. By these results, various thinning (and shrinking to a residue) algorithms can be verified. To illustrate the usefulness of our sufficient conditions, we propose a new parallel thinning algorithm and prove its topological correctness. [bibtex-key = NKP_IWCIA2011]


125. Jozsef Nemeth, Zoltan Kato, and Ian Jermyn. A Multi-Layer 'Gas of Circles' Markov Random Field Model for the Extraction of Overlapping Near-Circular Objects. In Jacques Blanc-Talon, Wilfried Philips, Dan Popescu, Paul Scheunders, and Richard Kleihorst, editors, Proceedings of the Advanced Concepts for Intelligent Vision Systems, volume 6915 of Lecture Notes in Computer Science, Ghent, Belgium, pages 171-182, August 2011. Springer Verlag. [PDF]
     Abstract: We propose a multi-layer binary Markov random field (MRF) model that assigns high probability to object configurations in the image domain consisting of an unknown number of possibly touching or overlapping near-circular objects of approximately a given size. Each layer has an associated binary field that specifies a region corresponding to objects. Overlapping objects are represented by regions in different layers. Within each layer, long-range interactions favor connected components of approximately circular shape, while regions in different layers that overlap are penalized. Used as a prior coupled with a suitable data likelihood, the model can be used for object extraction from images, e.g. cells in biological images or densely-packed tree crowns in remote sensing images. We present a theoretical and experimental analysis of the model, and demonstrate its performance on various synthetic and biomedical images. [bibtex-key = Nemeth-etal2011]


126. Gabor Nemeth and Kalman Palagyi. 2D parallel thinning algorithms based on isthmus-preservation. In S. Loncaric, G. Ramponi, and D. Sersic, editors, Proceedings of the International Symposium on Image and Signal Processing and Analysis (ISPA), Dubrovnik, Croatia, pages 585-590, September 2011. IEEE. [PDF]
     Abstract: Skeletons are widely used shape descriptors which summarize the general form of binary objects. A technique to obtain skeletons is the thinning, that is an iterative layer-by-layer erosion in a topology-preserving way. Conventional thinning algorithms preserve line endpoints to provide important geometric information relative to the object to be represented. Bertrand and Couprie proposed an alternative strategy by accumulating isthmus points that are line interior points. In this paper we present six new 2D parallel thinning algorithms that are derived from some sufficient conditions for topology preserving reductions and based on isthmus-preservation. [bibtex-key = NemethPalagyiISPA2011]


127. Attila Tanacs and Zoltan Kato. Fast linear registration of 3D objects segmented from medical images. In Proceedings of the International Conference on BioMedical Engineering and Informatics (BMEI), Shanghai, China, pages 299-303, October 2011. IEEE. [PDF]
     Abstract: In this paper a linear registration framework is used for medical image registration using segmented binary objects. The method is best suited for problems where the segmentation is available, but we also propose a general bone segmentation approach for CT images.We focus on the case when the objects to be registered differ considerably because of segmentation errors. We check the applicability of the method to bone segmentation of pelvic and thoracic CT images. Comparison is also made against a classical mutual information-based registration method. [bibtex-key = TanacsKatoBMEI2011]


128. Csaba Domokos and Zoltan Kato. Affine puzzle: Realigning deformed object fragments without correspondences. In Kostas Daniilidis, Petros Maragos, and Nikos Paragios, editors, Proceedings of the European Conference on Computer Vision, volume 6312 of Lecture Notes in Computer Science, Crete, Greece, pages 777-790, September 2010. Springer Verlag. [PDF]
     Abstract: This paper is addressing the problem of realigning broken objects without correspondences. We consider linear transformations between the object fragments and present the method through 2D and 3D affine transformations. The basic idea is to construct and solve a polynomial system of equations which provides the unknown parameters of the alignment. We have quantitatively evaluated the proposed algorithm on a large synthetic dataset containing 2D and 3D images. The results show that the method performs well and robust against segmentation errors. We also present experiments on 2D real images as well as on volumetric medical images applied to surgical planning. [bibtex-key = Domokos-Kato2010]


129. Norbert Hantos and Peter Balazs. Image Enhancement by Median Filters in Algebraic Reconstruction Methods: An Experimental Study. In R. Boyle, B. Parvin, D. Koracin, R. Chung, M. Hammoud; Hussain, K.-H. Tan, R. Crawfis, D. Thalmann, D. Kao, and L. Avila, editors, Proceedings of the International Symposium on Visual Computing, volume 6455 of Lecture Notes in Computer Science, Las Vegas, USA, pages 339-348, nov--dec 2010. Springer Verlag. [PDF]
     Abstract: Algebraic methods for image reconstruction provide good solutions even if only few projections are available. However, they can create noisy images if the number of iterations or the computational time is limited. In this paper, we show how to decrease the effect of noise by using median filters during the iterations. We present an extensive study by applying filters of different sizes and in various times of the reconstruction process. Also, our test images are of different structural complexity. Our study concentrates on the ART and its discrete variant DART reconstruction methods. [bibtex-key = HantosBalazs2010]


130. Peter Kardos, Gabor Nemeth, and Kalman Palagyi. Topology Preserving Parallel Smoothing for 3D Binary Images. In R.P. Barneva, V.E. Brimkov, H.A. Hauptman, R.M. Natal Jorge, and J.M.R.S. Tavares, editors, Proceedings of the International Symposium on Computational Modeling of Objects Represented in Images, volume 6026 of Lecture Notes in Computer Science, Buffalo, USA, pages 287-298, May 2010. Springer Verlag. [PDF] [doi:10.1007/978-3-642-12712-0_26]
     Abstract: This paper presents a new algorithm for smoothing 3D binary images in a topology preserving way. Our algorithm is a reduction operator: some border points that are considered as extremities are removed. The proposed method is composed of two parallel reduction operators. We are to apply our smoothing algorithm as an iteration-by-iteration pruning for reducing the noise sensitivity of 3D parallel surface-thinning algorithms. An efficient implementation of our algorithm is sketched and its topological correctness for (26,6) pictures is proved. [bibtex-key = NKP_CompIMAGE2010]


131. Milan Lesko, Zoltan Kato, Antal Nagy, Imre Gombos, Zsolt Torok, Laszlo Jr Vigh, and Laszlo Vigh. Live Cell Segmentation in Fluorescence Microscopy via Graph Cut. In Proceedings of the International Conference on Pattern Recognition, Istambul, Turkey, pages 1485-1488, August 2010. IAPR, IEEE. [PDF]
     Abstract: We propose a novel Markovian segmentation model which takes into account edge information. By construction, the model uses only pairwise interactions and its energy is submodular. Thus the exact energy minima is obtained via a max-flow/min-cut algorithm. The method has been quantitatively evaluated on synthetic images as well as on fluorescence microscopic images of live cells. [bibtex-key = Lesko-etal2010]


132. Gabor Nemeth, Peter Kardos, and Kalman Palagyi. Topology Preserving 2-Subfield 3D Thinning Algorithms. In Proceedings of the IASTED International Conference on Signal Processing, Pattern Recognition and Applications, Innsbruck, Austria, pages 310-316, February 2010. IASTED. [PDF]
     Abstract: This paper presents a new family of 3D thinning algorithms for extracting skeleton-like shape features (i.e, centerline, medial surface, and topological kernel) from volumetric images. A 2-subfield strategy is applied: all points in a 3D picture are partitioned into two subsets which are alternatively activated. At each iteration, a parallel operator is applied for deleting some border points in the active subfield. The proposed algorithms are derived from Ma's sufficient conditions for topology preservation, and they use various endpoint characterizations. [bibtex-key = NKP_SPPRA_2010]


133. Gabor Nemeth, Peter Kardos, and Kalman Palagyi. Topology Preserving 3D Thinning Algorithms Using Four and Eight Subfields. In Aurelio Campilho and Mohamed Kamel, editors, Proceedings of the International Conference on Image Analysis and Recognition, volume 6111 of Lecture Notes in Computer Science, Povoa de Varzim, Portugal, pages 316-325, June 2010. Springer Verlag. [PDF] [doi:10.1007/978-3-642-13772-3_32]
     Abstract: Thinning is a frequently applied technique for extracting skeleton-like shape features (i.e., centerline, medial surface, and topological kernel) from volumetric binary images. Subfield-based thinning algorithms partition the image into some subsets which are alternatively activated, and some points in the active subfield are deleted. This paper presents a set of new 3D parallel subfield-based thinning algorithms that use four and eight subfields. The three major contributions of this paper are: 1) The deletion rules of the presented algorithms are derived from some sufficient conditions for topology preservation. 2) A novel thinning scheme is proposed that uses iteration-level endpoint checking. 3) Various characterizations of endpoints yield different algorithms. [bibtex-key = NKP_ICIAR_2010]


134. Attila Tanacs, Natasa Sladoje, Joakim Lindblad, and Zoltan Kato. Estimation of linear deformations of 3D objects. In Proceedings of the International Conference on Image Processing, Hong Kong, China, pages 153-156, September 2010. IEEE, IEEE. [PDF]
     Abstract: We propose a registration method to find affine transformations between 3D objects by constructing and solving an overdetermined system of polynomial equations. We utilize voxel coverage information for more precise object boundary description. An iterative solution enables us to easily adjust the method to recover \eg rigid-body and similarity transformations. Synthetic tests show the advantage of the voxel coverage representation, and reveal the robustness properties of our method against different types of segmentation errors. The method is tested on a real medical CT volume.. [bibtex-key = Tanacs-etal2010]


135. Laszlo Varga, Peter Balazs, and Antal Nagy. Direction-dependency of a binary tomographic reconstruction algorithm. In R.P. Barneva, V.E. Brimkov, H.A. Hauptman, R.M. Natal Jorge, and J.M.R.S. Tavares, editors, Proceedings of the International Symposium on Computational Modeling of Objects Represented in Images, volume 6026 of Lecture Notes in Computer Science, Buffalo, USA, pages 242-253, May 2010. Springer Verlag. [PDF] [doi:10.1007/978-3-642-12712-0_22]
     Abstract: We study how the quality of an image reconstructed by a binary tomographic algorithm depends on the direction of the observed object in the scanner, if only a few projections are available. To do so we conduct experiments on a set of software phantoms by reconstructing them form different projection sets using an algorithm based on D.C. programming (a method for minimizing the difference of convex functions), and compare the accuracy of the corresponding reconstructions by two suitable approaches. Based on the experiments, we discuss consequences on applications arising from the field of non-destructive testing, as well. [bibtex-key = VargaBalazsNagy2010]


136. Laszlo Varga, Peter Balazs, and Antal Nagy. Projection Selection Algorithms for Discrete Tomography. In J. Blanc-Talon, D. Bone, W. Philips, D. Popescu, and P. Scheunders, editors, Proceedings of the Advanced Concepts for Intelligent Vision Systems, volume 6474 of Lecture Notes in Computer Science, Sydney, Australia, pages 390-401, December 2010. Springer Verlag. [PDF]
     Abstract: In this paper we study how the choice of projection angles affect the quality of the discrete tomographic reconstruction of an object. We supply four different strategies for selecting projection angle sets and compare them by conducting experiments on a set of software phantoms. We also discuss some consequences of our observations. Furthermore, we introduce a possible application of the proposed angle selection algorithms as well. [bibtex-key = ACIVS2010:Varga]


137. Peter Balazs. Reconstruction of canonical hv-convex discrete sets from horizontal and vertical projections. In Petra Wiederhold and Reneta P. Barneva, editors, Proceedings of the International Workshop on Combinatorial Image Analysis, volume 5852 of Lecture Notes in Computer Science, Playa del Carmen, Mexico, pages 280-288, November 2009. Springer Verlag. [PDF] [doi:10.1007/978-3-642-10210-3]
     Abstract: The problem of reconstructing some special hv-convex discrete sets from their two orthogonal projections is considered. In general, the problem is known to be NP-hard, but it is solvable in polynomial time if the discrete set to be reconstructed is also 8-connected. In this paper, we define an intermediate class - the class of hv-convex canonical discrete sets - and give a constructive proof that the above problem remains computationally tractable for this class, too. We also discuss some further theoretical consequences and present experimental results as well. [bibtex-key = Balazs2009iwcia]


138. Peter Balazs and Mihaly Gara. An Evolutionary Approach for Object-Based Image Reconstruction Using Learnt Priors. In Arnt-Borre Salberg, Jon Yngve Hardeberg, and Robert Jenssen, editors, Proceedings of the Scandinavian Conference on Image Analysis, volume 5575 of Lecture Notes in Computer Science, Oslo, Norway, pages 520-529, June 2009. Springer Verlag. [PDF]
     Abstract: In this paper we present a novel algorithm for reconstructing binary images containing objects which can be described by some parameters. In particular, we investigate the problem of reconstructing binary images representing disks from four projections. We develop a genetic algorithm for this and similar problems. We also discuss how prior information on the number of disks can be incorporated into the reconstruction in order to obtain more accurate images. In addition, we present a method to exploit such kind of knowledge from the projections themselves. Experiments on artificial data are also conducted. [bibtex-key = BalazsGara2009]


139. Tamas Blaskovics, Zoltan Kato, and Ian Jermyn. A Markov Random Field Model for Extracting Near-Circular Shapes. In Proceedings of the International Conference on Image Processing, Cairo, Egypt, pages 1073-1076, November 2009. IEEE, IEEE. [PDF]
     Abstract: We propose a binary Markov Random Field (MRF) model that assigns high probability to regions in the image domain consisting of an unknown number of circles of a given radius. We construct the model by discretizing the 'gas of circles' phase field model in a principled way, thereby creating an 'equivalent' MRF. The behaviour of the resulting MRF model is analyzed, and the performance of the new model is demonstrated on various synthetic images as well as on the problem of tree crown detection in aerial images. [bibtex-key = Blaskovics-etal2009a]


140. Csaba Domokos and Zoltan Kato. Affine Alignment of Compound Objects: A Direct Approach. In Proceedings of the International Conference on Image Processing, Cairo, Egypt, pages 169-172, November 2009. IEEE, IEEE. [PDF]
     Abstract: A direct approach for parametric estimation of 2D affine deformations between compound shapes is proposed. It provides the result as a least-square solution of a linear system of equations. The basic idea is to fit Gaussian densities over the objects yielding covariant functions, which preserves the effect of the unknown transformation. Based on these functions, linear equations are constructed by integrating nonlinear functions over appropriate domains. The main advantages are: linear complexity, easy implementation, works without any time consuming optimization or established correspondences. Comparative tests show that it outperforms state-ofthe- art methods both in terms of precision, robustness and complexity. [bibtex-key = Domokos-Kato2009b]


141. Balazs Erdohelyi and Endre Varga. Semi-automatic bone fracture reduction in surgical planning. In Proceedings of the International Conference on Computer Assisted Radiology and Surgery (CARS), volume 4 of International Journal of Computer Assisted Radiology and Surgery, Berlin, Germany, pages S98-S99, June 2009. Springer Verlag. [PDF] [doi:10.1007/s11548-009-0314-1]
     Abstract: The surgical intervention of complex bone fractures has to be planned very carefully, especially for such a complex region as the pelvic ring. The computer aided surgical planning is done before the actual surgery takes place and its main purpose is to gather more information about the dislocation of the bone fragments and to arrange the surgical implants to be inserted. With the help of finite element analysis even the biomechanical stability of the whole plan can be predicted. To create such a plan the following steps are performed. First, the CT dataset of the patient is segmented [1], which enables us to treat the different bones and broken fragments separately. Next the surface of this volumetric dataset is determined [2] and presented in a 3D environment. Since the fragments may have moved during the fracture, it is essential to move and rotate them back to their original anatomic position. Without this reduction, no implants can be inserted, since the final locations of the fragments are unknown. Previous solutions to the fracture reduction problem included moving the fragments with the mouse, or a special 3D haptic device. The former is not intuitive to use since the mouse is only 2D, and the later is expensive and still requires learning. [bibtex-key = ErdohelyiCARS2009]


142. Peter Kardos, Gabor Nemeth, and Kalman Palagyi. An Order-Independent Sequential Thinning Algorithm. In Petra Wiederhold and Reneta P. Barneva, editors, Proceedings of the International Workshop on Combinatorial Image Analysis, volume 5852 of Lecture Notes in Computer Science, Playa del Carmen, Mexico, pages 162-175, November 2009. Springer Verlag. [PDF]
     Abstract: Thinning is a widely used approach for skeletonization. Sequential thinning algorithms use contour tracking: they scan border points and remove the actual one if it is not designated a skeletal point. They may produce various skeletons for different visiting orders. In this paper, we present a new 2-dimensional sequential thinning algorithm, which produces the same result for arbitrary visiting orders and it is capable of extracting maximally thinned skeletons. [bibtex-key = KardosNemethPalagyi09]


143. Endre Katona. A Graph Based Data Model for Graphics Interpretation. In Andrea Torsello, Francisco Escolano Ruiz, and Luc Brun, editors, Proceedings of Graph-Based Representations in Pattern Recognition, volume 5534 of Lecture Notes in Computer Science, Venice, Italy, pages 355-364, October 2009. Springer Verlag. [WWW] [PDF]
     Abstract: A universal data model, named DG, is introduced to handle vectorized data uniformly during the whole recognition process. The model supports low level graph algorithms as well as higher level processing. To improve algorithmic efficiency, spatial indexing can be applied. Implementation aspects are discussed as well. An earlier version of the DG model has been applied for interpretation of Hungarian cadastral maps. Although this paper gives examples of map interpretation, our concept can be extended to other fields of graphics recognition. [bibtex-key = KatonaGBR2009]


144. Jozsef Nemeth, Csaba Domokos, and Zoltan Kato. Nonlinear Registration of Binary Shapes. In Proceedings of the International Conference on Image Processing, Cairo, Egypt, pages 1001-1004, November 2009. IEEE, IEEE. [PDF]
     Abstract: A novel approach is proposed to estimate the parameters of a diffeomorphism that aligns two binary images. Classical approaches usually define a cost function based on a similarity metric and then find the solution via optimization. Herein, we trace back the problem to the solution of a system of nonlinear equations which directly provides the parameters of the aligning transformation. The proposed method works without any time consuming optimization step or established correspondences. The advantage of our algorithm is that it is easy to implement, less sensitive to the strength of the deformation, and robust against segmentation errors. The efficiency of the proposed approach has been demonstrated on a large synthetic dataset as well as in the context of an industrial application. [bibtex-key = Nemeth-etal2009b]


145. Jozsef Nemeth, Csaba Domokos, and Zoltan Kato. Recovering Planar Homographies between 2D Shapes. In Proceedings of the International Conference on Computer Vision, Kyoto, Japan, pages 2170-2176, September 2009. IEEE, IEEE. [PDF]
     Abstract: Images taken from different views of a planar object are related by planar homography. Recovering the parameters of such transformations is a fundamental problem in computer vision with various applications. This paper proposes a novel method to estimate the parameters of a homography that aligns two binary images. It is obtained by solving a system of nonlinear equations generated by integrating linearly independent functions over the domains determined by the shapes. The advantage of the proposed solution is that it is easy to implement, less sensitive to the strength of the deformation, works without established correspondences and robust against segmentation errors. The method has been tested on synthetic as well as on real images and its efficiency has been demonstrated in the context of two different applications: alignment of hip prosthesis X-ray images and matching of traffic signs. [bibtex-key = Nemeth-etal2009a]


146. Laszlo G. Nyul. Retinal image analysis for automated glaucoma risk evaluation. In Jianguo Liu, Kunio Doi, Aaron Fenster, and S. C. Chan, editors, Proceedings of MIPPR 2009: Medical Imaging, Parallel Processing of Images, and Optimization Techniques, volume 7497 of SPIE Proceedings, Yichang, China, pages 74971C1-9, October 2009. SPIE. [WWW] [PDF] [doi:10.1117/12.851179]
     Abstract: Images of the eye ground not only provide an insight to important parts of the visual system but also reflect the general state of health of the entire human body. Automatic retina image analysis is becoming an important screening tool for early detection of certain risks and diseases. Glaucoma is one of the most common causes of blindness and is becoming even more important considering the ageing society. Robust mass-screening may help to extend the symptom-free life of affected patients. Our research is focused on a novel automated classification system for glaucoma, based on image features from fundus photographs. Our new data-driven approach requires no manual assistance and does not depend on explicit structure segmentation and measurements. First, disease independent variations, such as nonuniform illumination, size differences, and blood vessels are eliminated from the images. Then, the extracted high-dimensional feature vectors are compressed via PCA and combined before classification with SVMs takes place. The technique achieves an accuracy of detecting glaucomatous retina fundus images comparable to that of human experts. The "vessel-free" images and intermediate output of the methods are novel representations of the data for the physicians that may provide new insight into and help to better understand glaucoma. [bibtex-key = nyul:74971C]


147. Kalman Palagyi and Gabor Nemeth. Fully Parallel 3D Thinning Algorithms Based on Sufficient Conditions for Topology Preservation. In Srecko Brlek, Christophe Reutenauer, and Xavier Provencal, editors, Proceedings of the International Conference on Discrete Geometry for Computer Imagery, volume 5810 of Lecture Notes in Computer Science, Montreal, Quebec, Canada, pages 481-492, sep--oct 2009. Springer Verlag. [PDF]
     Abstract: This paper presents a family of parallel thinning algorithms for extracting medial surfaces from 3D binary pictures. The proposed algorithms are based on sufficient conditions for 3D parallel reduction operators to preserve topology for (26,6) pictures. Hence it is self-evident that our algorithms are topology preserving. Their efficient implementation on conventional sequential computers is also presented. [bibtex-key = PalagyiNemeth09]


148. Attila Tanacs, Csaba Domokos, Natasa Sladoje, Joakim Lindblad, and Zoltan Kato. Recovering affine deformations of fuzzy shapes. In Arnt-Borre Salberg, Jon Yngve Hardeberg, and Robert Jenssen, editors, Proceedings of the Scandinavian Conference on Image Analysis, volume 5575 of Lecture Notes in Computer Science, Oslo, Norway, pages 735-744, June 2009. Springer Verlag. [PDF]
     Abstract: Fuzzy sets and fuzzy techniques are attracting increasing attention nowadays in the field of image processing and analysis. It has been shown that the information preserved by using fuzzy representation based on area coverage may be successfully utilized to improve precision and accuracy of several shape descriptors; geometric moments of a shape are among them. We propose to extend an existing binary shape matching method to take advantage of fuzzy object representation. The result of a synthetic test show that fuzzy representation yields smaller registration errors in average. A segmentation method is also presented to generate fuzzy segmentations of real images. The applicability of the proposed methods is demonstrated on real X-ray images of hip replacement implants. [bibtex-key = Tanacs:2009:SCIA]


149. Marton Balasko, Attila Kuba, Attila Tanacs, Zoltan Kiss, Antal Nagy, and Burkhard Schillinger. Comparison Radiography and Tomography Possibilities of FRM-II (20 MW) and Budapest (10 MW) Research Reactor. In Muhammed Arif, editor, Proceedings of the Eight World Conference WCNR-8, pages 18-27, October 2008. [PDF]
     Abstract: In November of 2005 a measurement series was conducted on different kinds of reference objects to compare the results of the FRM-II (ANTARES) and the Budapest (Radiography Station) research reactor in the field of neutron radiography, gamma radiography, classical tomography and discrete tomography by an ANDOR CCD camera and Imaging plate with BAS 2500 Scanner (contributed by AIDA software). [bibtex-key = BalaskoKuba:2008:Comparison]


150. Marton Balasko, Erzsebet Svab, Zoltan Kiss, Attila Tanacs, Antal Nagy, and Attila Kuba. Study of the Inner Structure of a Damaged Control Rod by Neutron and X-ray Radiography and Discrete Tomography. In Muhammed Arif, editor, Proceedings of the Eight World Conference WCNR-8, pages 294-303, October 2008. [PDF]
     Abstract: A control rod K5 of the 10 MW Budapest Research Reactor was damaged in 2004. We have applied both neutron and X-ray radiography to study the distribution of the residual material in the damaged K5. X-ray radiography visualized the main distribution of the residual materials, like different kinds of B4C and water, and the frozen gas bubbles, while neutron radiography displayed the fine structure of the congealed B4C. The 3D reconstruction of the most interesting parts of K5 has been done by discrete tomography. [bibtex-key = BalaskoKiss:2008:Study]


151. Peter Balazs. On the number hv-convex discrete sets. In Valentin E. Brimkov, Reneta P. Barneva, and Herbert A. Hauptman, editors, Proceedings of the International Workshop on Combinatorial Image Analysis, volume 4958 of Lecture Notes in Computer Science, Buffalo, USA, pages 112-123, April 2008. Springer Verlag. [PDF]
     Abstract: One of the basic problems in discrete tomography is the reconstruction of discrete sets from few projections. Assuming that the set to be reconstructed fulfills some geometrical properties is a commonly used technique to reduce the number of possibly many different solutions of the same reconstruction problem. The class of $hv$-convex discrete sets and its subclasses have a well-developed theory. Several reconstruction algorithms as well as some complexity results are known for those classes. The key to achieve polynomial-time reconstruction of an $hv$-convex discrete set is to have the additional assumption that the set is connected as well. This paper collects several statistics on $hv$-convex discrete sets, which are of great importance in the analysis of algorithms for reconstructing such kind of discrete sets. [bibtex-key = Balazs2008a]


152. Peter Balazs. Reconstruction of binary images with few disjoint components from two projections. In George Bebis et al., editor, Proceedings of the International Symposium on Visual Computing, volume 5359 of Lecture Notes in Computer Science, Las Vegas, NV, USA, pages 1147-1156, December 2008. Springer Verlag. [PDF]
     Abstract: We present a general framework for reconstructing binary images with few disjoint components from the horizontal and vertical projections. We develop a backtracking algorithm that works for binary images having components from an arbitrary class. Thus, a priori information about the components of the image to be reconstructed can be incorporated into the reconstruction process. In addition, we can keep control over the number of components which can increase the speed and accuracy of the reconstruction. Experimental results are also presented. [bibtex-key = Balazs2008c]


153. Peter Balazs and Mihaly Gara. Decision trees in binary tomography for supporting the reconstruction of hv-convex connected images. In S. Bourennane, W. Philips, D. Popescu, and P. Scheunders, editors, Proceedings of the Advanced Concepts for Intelligent Vision Systems, volume 5259 of Lecture Notes in Computer Science, Juan-les-Pins, France, pages 433-443, October 2008. Springer Verlag. [PDF]
     Abstract: In binary tomography, several algorithms are known for reconstructing binary images having some geometrical properties from their projections. In order to choose the appropriate reconstruction algorithm it is necessary to have a priori information of the image to be reconstructed. In this way we can improve the speed and reduce the ambiguity of the reconstruction. Our work is concerned with the problem of retrieving geometrical information from the projections themselves. We investigate whether it is possible to determine geometric features of binary images if only their projections are known. Most of the reconstruction algorithms based on geometrical information suppose $hv$-convexity or connectedness about the image to be reconstructed. We investigate those properties in detail, and also the task of separating 4- and 8-connected images. We suggest decision trees for the classification, and show some preliminary experimental results of applying them for the class of $hv$-convex and connected discrete sets. [bibtex-key = BalazsGara2008]


154. Csaba Domokos and Zoltan Kato. Binary Image Registration Using Covariant Gaussian Densities. In A. Campilho and M. Kamel, editors, Proceedings of the International Conference on Image Analysis and Recognition, volume 5112 of Lecture Notes in Computer Science, Povoa de Varzim, Portugal, pages 455-464, June 2008. Springer Verlag. [PDF]
     Abstract: We consider the estimation of 2D affine transformations aligning a known binary shape and its distorted observation. The classical way to solve this registration problem is to find correspondences between the two images and then compute the transformation parameters from these landmarks. In this paper, we propose a novel approach where the exact transformation is obtained as a least-squares solution of a linear system. The basic idea is to fit a Gaussian density to the shapes which preserves the effect of the unknown transformation. It can also be regarded as a consistent coloring of the shapes yielding two rich functions defined over the two shapes to be matched. The advantage of the proposed solution is that it is fast, easy to implement, works without established correspondences and provides a unique and exact solution regardless of the magnitude of transformation. [bibtex-key = Domokos-Kato2008]


155. Csaba Domokos, Zoltan Kato, and Joseph M. Francos. Parametric Estimation of Affine Deformations of Binary Images. In Proceedings of the International Conference on Acoustics, Speech and Signal Processing, Las Vegas, Nevada, USA, pages 889-892, April 2008. IEEE, IEEE. [PDF] [doi:10.1109/ICASSP.2008.4517753]
     Abstract: We consider the problem of planar object registration on binary images where the aligning transformation is restricted to the group of {\em affine} transformations. Previous approaches usually require established correspondences or the solution of nonlinear optimization problems. Herein we show that it is possible to formulate the problem as the solution of a system of up to third order polynomial equations. These equations are constructed in a simple way using some basic geometric information of {\em binary} images. It does not need established correspondences nor the solution of complex optimization problems. The resulting algorithm is fast and provides a direct solution regardless of the magnitude of transformation. [bibtex-key = Domokos-etal2008]


156. Attila Fazekas, Kalman Palagyi, Gyorgy Kovacs, and Gabor Nemeth. Skeletonization Based on Metrical Neighborhood Sequences. In Antonios Gasteratos, Markus Vincze, and John K. Tsotsos, editors, Proceedings of the International Conference on Computer Vision Systems, volume 5008 of Lecture Notes in Computer Science, Santorini, Greece, pages 333-342, May 2008. Springer Verlag. [PDF]
     Abstract: Skeleton is a shape descriptor which summarizes the general form of objects. It can be expressed in terms of the fundamental morphological operations. The limitation of that characterization is that its construction based on digital disks such that cannot provide good approximation to the Euclidean disks. In this paper we define a new type of skeleton based on neighborhood sequences that is much closer to the Euclidean skeleton. A novel method for quantitative comparison of skeletonization algorithms is also proposed. [bibtex-key = FazekasEtalICVS2008]


157. Peter Balazs. Generation and empirical investigation of hv-convex discrete sets. In Bjarne K. Ersboll and Kim S. Pedersen, editors, Proceedings of the Scandinavian Conference on Image Analysis, volume 4522 of Lecture Notes in Computer Science, Aalborg, Denmark, pages 344-353, June 2007. Springer Verlag. [PDF]
     Abstract: One of the basic problems in discrete tomography is the reconstruction of discrete sets from few projections. Assuming that the set to be reconstructed fulfils some geometrical properties is a commonly used technique to reduce the number of possibly many different solutions of the same reconstruction problem. Since the reconstruction from two projections in the class of so-called $hv$-convex sets is NP-hard this class is suitable to test the efficiency of newly developed reconstruction algorithms. However, until now no method was known to generate sets of this class from uniform random distribution and thus only ad hoc comparison of several reconstruction techniques was possible. In this paper we first describe a method to generate some special $hv$-convex discrete sets from uniform random distribution. Moreover, we show that the developed generation technique can easily be adapted to other classes of discrete sets, even for the whole class of $hv$-convexes. Several statistics are also presented which are of great importance in the analysis of algorithms for reconstructing $hv$-convex sets. [bibtex-key = Balazs2007]


158. Gyorgy Bekes, Marta Fidrich, Laszlo G. Nyul, Attila Kuba, and Eors Mate. 3D segmentation of liver, kidneys and spleen from CT images. In Proceedings of the International Conference on Computer Assisted Radiology and Surgery (CARS), volume 2 of International Journal of Computer Assisted Radiology and Surgery, Berlin, Germany, pages S45-S47, June 2007. Springer Verlag. [WWW] [PDF] [doi:10.1007/s11548-007-0083-7]
     Abstract: The clinicians often need to segment the abdominal organs for radiotherapy planning. Manual segmentation of these organs is very time-consuming, therefore automated methods are desired. We developed a semi-automatic segmentation method to outline liver, spleen and kidneys. It works on CT images without contrast intake that are acquired with a routine clinical protocol. From an initial surface around a user defined seed point, the segmentation of the organ is obtained by an active surface algorithm. Pre- and post-processing steps are used to adapt the general method for specific organs. The evaluation results show that the accuracy of our method is about 90%, which can be further improved with little manual editing, and that the precision is slightly higher than that of manual contouring. Our method is accurate, precise and fast enough to use in the clinical practice. [bibtex-key = Bekes2007]


159. Csaba Benedek, Tamas Sziranyi, Zoltan Kato, and Josiane Zerubia. A Multi-Layer MRF Model for Object-Motion Detection in Uregistered Airborne Image-Pairs. In Proceedings of the International Conference on Image Processing, volume VI, San Antonio, Texas, USA, pages 141-144, September 2007. IEEE, IEEE. [PDF]
     Abstract: In this paper, we give a probabilistic model for automatic change detection on airborne images taken with moving cameras. To ensure robustness, we adopt an unsupervised coarse matching instead of a precise image registration. The challenge of the proposed model is to eliminate the registration errors, noise and the parallax artifacts caused by the static objects having considerable height (buildings, trees, walls etc.) from the difference image. We describe the background membership of a given image point through two different features, and introduce a novel three-layer Markov Random Field (MRF) model to ensure connected homogenous regions in the segmented image. [bibtex-key = Benedek-etal2007]


160. Rudiger Bock, Jorg Meier, Georg Michelson, Laszlo G. Nyul, and Joachim Hornegger. Classifying Glaucoma with Image-based Features from Fundus Photographs. In F. A. Hamprect, C. Schnorr, and B. Jähne, editors, Proceedings of the Annual Symposium of the German Association for Pattern Recognition (DAGM), volume 4713 of Lecture Notes in Computer Science, Heidelberg, Germany, pages 355-364, 2007. Springer Verlag. [PDF]
     Abstract: Glaucoma is one of the most common causes of blindness and it is becoming even more important considering the ageing society. Because healing of died retinal nerve fibers is not possible early detection and prevention is essential. Robust, automated mass-screening will help to extend the symptom-free life of affected patients. We devised a novel, automated, appearance based glaucoma classification system that does not depend on segmentation based measurements. Our purely data-driven approach is applicable in large-scale screening examinations. It applies a standard pattern recognition pipeline with a 2-stage classification step. Several types of image-based features were analyzed and are combined to capture glaucomatous structures. Certain disease independent variations such as illumination inhomogeneities, size differences, and vessel structures are eliminated in the preprocessing phase. The ``vesselfree'' images and intermediate results of the methods are novel representations of the data for the physicians that may provide new insight into and help to better understand glaucoma. Our system achieves 86% success rate on a data set containing a mixture of 200 real images of healthy and glaucomatous eyes. The performance of the system is comparable to human medical experts in detecting glaucomatous retina fundus images. [bibtex-key = Bock:2007:CGI]


161. Balazs Erdohelyi, Endre Varga, and Attila Kuba. Surgical Planning Tool with Biomechanical Simulation. In Proceedings of the International Conference on Computer Assisted Radiology and Surgery (CARS), volume 2 of International Journal of Computer Assisted Radiology and Surgery, Berlin, Germany, pages S262-S263, 2007. [PDF] [doi:10.1007/s11548-007-0098-0]
     Abstract: The fixation of fractured bones often needs a very careful decision making. The operation has to be designed very carefully. A computer assisted system can help the surgeon in the planning phase to increase surgical accuracy. This paper introduces a software tool to plan a surgical intervention and to calculate the biomechanical stability of the plan. The proposed system provides 3D movement and rotation of the bone fragments and the insertion of fixation screws and plates. Finite element analysis is used to calculate mechanical stability of the surgical plan. Using these results the surgeon is able to see the week points of the fixation before the surgery. He can even try several surgical plans to pick the most promising one. [bibtex-key = Erdohelyi2007]


162. Peter Horvath. A Multispectral Data Model for Higher-Order Active Contours and its Application to Tree Crown Extraction. In Wilfried Philips, Dan Popescu, and Paul Scheunders, editors, Proceedings of the Advanced Concepts for Intelligent Vision Systems, volume 4678 of Lecture Notes in Computer Science, Delft, Netherlands, pages 200-211, August 2007. [PDF] [doi:10.1007/978-3-540-74607-2_18]
     Abstract: Forestry management makes great use of statistics concerning the individual trees making up a forest, but the acquisition of this information is expensive. Image processing can potentially both reduce this cost and improve the statistics. The key problem is the delineation of tree crowns in aerial images. The automatic solution of this problem requires considerable prior information to be built into the image and region models. Our previous work has focused on including shape information in the region model; in this paper we examine the image model. The aerial images involved have three bands. We study the statistics of these bands, and construct both multispectral and single band image models. We combine these with a higher-order active contour model of a `gas of circles' in order to include prior shape information about the region occupied by the tree crowns in the image domain. We compare the results produced by these models on real aerial images and conclude that multiple bands improves the quality of the segmentation. The model has many other potential applications, e.g. to nano-technology, microbiology, physics, and medical imaging. [bibtex-key = Horvath07c]


163. Peter Horvath and Ian Jermyn. A 'gas of Circles' Phase Field Model and its Application to Tree Crown Extraction. In Marek Domanski, Ryszard Stasinski, and Maciej Bartkowiak, editors, Proceedings of the European Signal Processing Conference (EUSIPCO), Poznan, Poland, September 2007. [PDF]
     Abstract: The problem of extracting the region in the image domain corresponding to an a priori unknown number of circular objects occurs in several domains. We propose a new model of a `gas of circles', the ensemble of regions in the image domain composed of circles of a given radius. The model uses the phase field reformulation of higher-order active contours (HOACs). Phase fields possess several advantages over contour and level set approaches to region modelling, in particular for HOAC models. The reformulation allows us to benefit from these advantages without losing the strengths of the HOAC framework. Combined with a suitable likelihood energy, and applied to the tree crown extraction problem, the new model shows markedly improved performance, both in quality of results and in computation time, which is two orders of magnitude less than the HOAC level set implementation. [bibtex-key = Horvath07d]


164. Peter Horvath and Ian Jermyn. A New Phase Field Model of a 'gas of Circles' for Tree Crown Extraction from Aerial Images. In Walter G. Kropatsch, Martin Kampel, and Allan Hanbury, editors, Proceedings of the International Conference on Computer Analysis of Images and Patterns, volume 4673 of Lecture Notes in Computer Science, Vienna, Austria, pages 702-709, August 2007. [PDF] [doi:10.1007/978-3-540-74272-2_87]
     Abstract: We describe a model for tree crown extraction from aerial images, a problem of great practical importance for the forestry industry. The novelty lies in the prior model of the region occupied by tree crowns in the image, which is a phase field version of the higher-order active contour inflection point `gas of circles' model. The model combines the strengths of the inflection point model with those of the phase field framework: it removes the `phantom circles' produced by the original `gas of circles' model, while executing two orders of magnitude faster than the contour-based inflection point model. The model has many other areas of application e.g., to imagery in nanotechnology, biology, and physics. [bibtex-key = Horvath07b]


165. Jorg Meier, Rudiger Bock, Georg Michelson, Laszlo G. Nyul, and Joachim Hornegger. Effects of Preprocessing Eye Fundus Images on Appearance Based Glaucoma Classification. In W. G. Kropatsch, M. Kampel, and A. Hanbury, editors, Proceedings of the International Conference on Computer Analysis of Images and Patterns, volume 4673 of Lecture Notes in Computer Science, Vienna, Austria, pages 165-172, 2007. Springer Verlag. [PDF]
     Abstract: Early detection of glaucoma is essential for preventing one of the most common causes of blindness. Our research is focused on a novel automated classification system based on image features from fundus photographs which does not depend on structure segmentation or prior expert knowledge. Our new data driven approach that needs no manual assistance achieves an accuracy of detecting glaucomatous retina fundus images compareable to human experts. In this paper, we study image preprocessing methods to provide better input for more reliable automated glaucoma detection. We reduce disease independent variations without removing information that discriminates between images of healthy and glaucomatous eyes. In particular, nonuniform illumination is corrected, blood vessels are inpainted and the region of interest is normalized before feature extraction and subsequent classification. The effect of these steps was evaluated using principal component analysis for dimension reduction and support vector machine as classifier. [bibtex-key = Meier:2007:EPE]


166. Kalman Palagyi. A 3-Subiteration Surface-Thinning Algorithm. In Walter G. Kropatsch, Martin Kampel, and Allan Hanbury, editors, Proceedings of the International Conference on Computer Analysis of Images and Patterns, volume 4673 of Lecture Notes in Computer Science, Vienna, Austria, pages 628-635, August 2007. Springer Verlag. [PDF]
     Abstract: Thinning is an iterative layer by layer erosion for extracting skeleton. This paper presents an efficient parallel 3D thinning algorithm which produces medial surfaces. A three-subiteration strategy is proposed: the thinning operation is changed from iteration to iteration with a period of three according to the three deletion directions. [bibtex-key = PalagyiCAIP2007]


167. Kalman Palagyi. A Subiteration-Based Surface-Thinning Algorithm with a Period of Three. In Fred A. Hamprecht, Christoph Schnorr, and Bernd Jähne, editors, Proceedings of the Annual Symposium of the German Association for Pattern Recognition (DAGM), volume 4713 of Lecture Notes in Computer Science, Heidelberg, Germany, pages 294-303, September 2007. Springer Verlag. [PDF]
     Abstract: Thinning on binary images is an iterative layer by layer erosion until only the "skeletons" of the objects are left. This paper presents an efficient parallel 3D surface-thinning algorithm. A three-subiteration strategy is proposed: the thinning operation is changed from iteration to iteration with a period of three according to the three deletion directions. [bibtex-key = PalagyiDAGM2007]


168. Peter Balazs. The number of line-convex directed polyominoes having the same orthogonal projections. In Attila Kuba, Laszlo G. Nyul, and Kalman Palagyi, editors, Proceedings of the International Conference on Discrete Geometry for Computer Imagery, volume 4245 of Lecture Notes in Computer Science, Szeged, Hungary, pages 77-85, October 2006. Springer Verlag. [PDF]
     Abstract: The number of line-convex directed polyominoes with given horizontal and vertical projections is studied. It is proven that diagonally convex directed polyominoes are uniquely determined by their orthogonal projections. The proof of this result is algorithmical. As a counterpart, we show that ambiguity can be exponential if antidiagonal convexity is assumed about the polyomino. Then, the results are generalised to polyominoes having convexity property along arbitrary lines. [bibtex-key = Balazs2006]


169. Sara Brunetti, Alain Daurat, and Attila Kuba. Fast Filling Operations Used in the Reconstruction of Convex Lattice Sets. In Attila Kuba, Laszlo G. Nyul, and Kalman Palagyi, editors, Proceedings of the International Conference on Discrete Geometry for Computer Imagery, volume 4245 of Lecture Notes in Computer Science, Szeged, Hungary, pages 98-109, October 2006. Springer Verlag. [PDF] [doi:10.1007/11907350_9]
     Abstract: Filling operations are procedures which are used in Discrete Tomography for the reconstruction of lattice sets having some convexity constraints. In [1], an algorithm which performs four of these filling operations has a time complexity of O(N2 logN), where N is the size of projections, and leads to a reconstruction algorithm for convex polyominoes running in O(N6 logN)-time. In this paper we first improve the implementation of these four filling operations to a time complexity of O(N2), and additionally we provide an implementation of a fifth filling operation (introduced in [2]) in O(N2 logN) that permits to decrease the overall time-complexity of the reconstruction algorithm to O(N4 logN). More generally, the reconstruction of Q-convex sets and convex lattice sets (intersection of a convex polygon with Z2) can be done in O(N4 logN)-time. [bibtex-key = Brunetti2006]


170. Peter Horvath, Ian Jermyn, Zoltan Kato, and Josiane Zerubia. A Higher-Order Active Contour Model for Tree Detection. In Proceedings of the International Conference on Pattern Recognition, volume 2, Hong Kong, China, pages 130-133, August 2006. IAPR, IEEE. [PDF]
     Abstract: We present a model of a 'gas of circles', the ensemble of regions in the image domain consisting of an unknown number of circles with approximately fixed radius and short range repulsive interactions, and apply it to the extraction of tree crowns from aerial images. The method uses the recently introduced 'higher order active contours' (HOACs), which incorporate long-range interactions between contour points, and thereby include prior geometric information without using a template shape. This makes them ideal when looking for multiple instances of an entity in an image. We study an existing HOAC model for networks, and show via a stability calculation that circles stable to perturbations are possible for constrained parameter sets. Combining this prior energy with a data term, we show results on aerial imagery that demonstrate the effectiveness of the method and the need for prior geometric knowledge. The model has many other potential applications. [bibtex-key = Horvath-etal2006]


171. Peter Horvath, Ian Jermyn, Zoltan Kato, and Josiane Zerubia. An improved `gas of circles' higher-order active contour model and its application to tree crown extraction. In Prem Kalra and Shmuel Peleg, editors, Proceedings of the Indian Conference on Computer Vision, Graphics and Image Processing, volume 4338 of Lecture Notes in Computer Science, Madurai, India, pages 152-161, December 2006. Springer Verlag. [PDF]
     Abstract: A central task in image processing is to find the region in the image corresponding to an entity. In a number of problems, the region takes the form of a collection of circles, e.g. tree crowns in remote sensing imagery; cells in biological and medical imagery. In [1], a model of such regions, the 'gas of circles' model, was developed based on higher-order active contours, a recently developed framework for the inclusion of prior knowledge in active contour energies. However, the model suffers from a defect. In [1], the model parameters were adjusted so that the circles were local energy minima. Gradient descent can become stuck in these minima, producing phantom circles even with no supporting data. We solve this problem by calculating, via a Taylor expansion of the energy, parameter values that make circles into energy inflection points rather than minima. As a bonus, the constraint halves the number of model parameters, and severely constrains one of the two that remain, a major advantage for an energy-based model. We use the model for tree crown extraction from aerial images. Experiments show that despite the lack of parametric freedom, the new model performs better than the old, and much better than a classical active contour. [bibtex-key = Horvath-etal2006a]


172. Zoltan Kato and Ting Chuen Pong. A Multi-Layer MRF Model for Video Object Segmentation. In P. J. Narayanan, Shree K. Nayar, and Heung-Yeung Shum, editors, Proceedings of the Asian Conference on Computer Vision, volume 3852 of Lecture Notes in Computer Science, Hyderabad, India, pages 953-962, January 2006. Springer Verlag. [PDF]
     Abstract: A novel video object segmentation method is proposed which aims at combining color and motion information. The model has a multilayer structure: Each feature has its own layer, called feature layer, where a classical Markov random field (MRF) image segmentation model is defined using only the corresponding feature. A special layer is assigned to the combined MRF model, called combined layer, which interacts with each feature layer and provides the segmentation based on the combination of different features. Unlike previous methods, our approach doesn’t assume motion boundaries being part of spatial ones. Therefore a very important property of the proposed method is the ability to detect boundaries that are visible only in the motion feature as well as those visible only in the color one. The method is validated on synthetic and real video sequences. [bibtex-key = Kato-Pong2006]


173. Stefan Weber, Antal Nagy, Thomas Schule, Christoph Schnorr, and Attila Kuba. A Benchmark Evaluation of Large-Scale Optimization Approaches to Binary Tomography. In Attila Kuba, Laszlo G. Nyul, and Kalman Palagyi, editors, Proceedings of the International Conference on Discrete Geometry for Computer Imagery, volume 4245 of Lecture Notes in Computer Science, Szeged, Hungary, pages 146-156, October 2006. Springer Verlag. [PDF] [doi:10.1007/11907350_13]
     Abstract: Discrete tomography concerns the reconstruction of functions with a finite number of values from few projections. For a number of important real-world problems, this tomography problem involves thousands of variables. Applicability and performance of discrete tomography therefore largely depend on the criteria used for reconstruction and the optimization algorithm applied. From this viewpoint, we evaluate two major optimization strategies, simulated annealing and convex-concave regularization, for the case of binary-valued functions using various data sets. Extensive numerical experiments show that despite being quite different from the viewpoint of optimization, both strategies show similar reconstruction performance as well as robustness to noise. [bibtex-key = Weber2006]


174. Stefan Weber, Thomas Schule, Attila Kuba, and Christoph Schnorr. Binary Tomography with Deblurring. In R. Reulke, U. Eckardt, B. Flach, U. Knauer, and K. Polthier, editors, Proceedings of the International Workshop on Combinatorial Image Analysis, volume 4040 of Lecture Notes in Computer Science, Berlin, Germany, pages 375-388, June 2006. Springer Verlag. [WWW] [doi:10.1007/11774938]
     Abstract: We study two scenarios of limited-angle binary tomography with data distorted with an unknown convolution: Either the projection data are taken from a blurred object, or the projection data themselves are blurred. These scenarios are relevant in case of scattering and due to a finite resolution of the detectors. Assuming that the unknown blurring process is adequately modeled by an isotropic Gaussian convolution kernel with unknown scale-parameter, we show that parameter estimation can be combined with the reconstruction process. To this end, a recently introduced Difference-of-Convex-Functions programming approach to limited-angle binary tomographic reconstruction is complemented with Expectation-Maximization iteration. Experimental results show that the resulting approach is able to cope with both ill-posed problems, limited-angle reconstruction and deblurring, simultaneously. [bibtex-key = Weber2006a]


175. Marton Balasko, Attila Kuba, Antal Nagy, Zoltan Kiss, Lajos Rodek, and Laszlo Rusko. Neutron-, gamma- and X-ray three-dimensional computed tomography at the Budapest research reactor site. In Proceedings of the International Topical Meeting on Neutron Radiography, volume 542 of Nuclear Instruments and Methods in Physics Research, Garching, Germany, pages 22-27, April 2005. [PDF] [doi:10.1016/j.nima.2005.01.006]
     Abstract: A new complex, neutron-, gamma- and X-ray three-dimensional computer tomography system suitable for experimental and industrial applications has been built at the 10-MW Budapest research reactor site. After the system was installed, a number of objects were investigated and tomographic projections were made. The evaluation relied on two reconstruction approaches. One of these is the classical, filtered back-projection method using 180 projected pictures, while the other is based on discrete tomography optimization algorithms where fewer projections were needed. [bibtex-key = Balasko2005]


176. Marton Balasko, Erzsebet Svab, Attila Kuba, Zoltan Kiss, Lajos Rodek, and Antal Nagy. Pipe corrosion and deposit study using neutron- and gamma- radiation sources. In Proceedings of the International Topical Meeting on Neutron Radiography, volume 542 of Nuclear Instruments and Methods in Physics Research, Garching, Germany, pages 302-308, April 2005. [PDF] [doi:10.1016/j.nima.2005.01.153]
     Abstract: The problems of corrosion and deposit are crucial issues in the pipelines of the chemical, nuclear and petrochemical industries. Radiography (neutron, gamma, X-ray) has long been used as a technique for pipe inspection and corrosion monitoring. The 10 MW Budapest research reactor site is a source of various energy neutron (thermal and epithermal) and gamma radiation. The detector system was a Peltier-cooled LLL CCD camera controlled by a PC with Image ProLite software and imaging plate equipment with a BAS 2500 scanner that used AIDA software. The objects inspected were corroded tubes and various kinds of test specimens with a large wall thickness (25 mm) inside and outside steps. In the evaluation part we used tomographic algorithms. A software simulation study was made as well. Fan-beam projections were computed of the given software phantoms and a new discrete tomography method was used to reconstruct the unknown objects from these projections. [bibtex-key = Balasko2005a]


177. Peter Balazs. Reconstruction of decomposable discrete sets from four projections. In Eric Andres, Guillaume Damiand, and Pascal Lienhardt, editors, Proceedings of the International Conference on Discrete Geometry for Computer Imagery, volume 3429 of Lecture Notes in Computer Science, Poitiers, France, pages 104-114, April 2005. Springer Verlag. [PDF]
     Abstract: In this paper we introduce the class of decomposable discrete sets and give a polynomial algorithm for reconstructing discrete sets of this class from four projections. It is also shown that the class of decomposable discrete sets is more general than the class $\mathcal S'_8$ of $hv$-convex 8- but not 4-connected discrete sets which was. As a consequence we also get that the reconstruction from four projections in ${\mathcal S}'_8$ can be solved in $O(mn)$ time. [bibtex-key = Balazs2005]


178. Peter Balazs. Reconstruction of discrete sets from four projections: strong decomposability. In Proceedings of the Workshop on Discrete Tomography and its Applications, volume 20 of Electronic Notes in Discrete Mathematics, New York City, USA, pages 329-345, June 2005. Elsevier. [PDF]
     Abstract: In this paper we introduce the class of strongly decomposable discrete sets and give an efficient algorithm for reconstructing discrete sets of this class from four projections. It is also shown that every Q-convex set (along the set of directions $\{ x,y\}$) consisting of several components is strongly decomposable. As a consequence of strong decomposability we get that in a subclass of $hv$-convex discrete sets the reconstruction from four projections can be solved in polynomial time. [bibtex-key = Balazs2005a]


179. Reinhard Beichel, Thomas Pock, Christian Janko, Roman Zotter, Bernhard Reitinger, Alexander Bornik, Kalman Palagyi, Erich Sorantin, Georg Werkgartner, Horst Bischof, and Milan Sonka. Liver segment approximation in CT data for surgical resection planning. In Fitzpatrick J. M. and Sonka M., editors, Proceedings of Medical Imaging 2004: Image Processing, volume 5370 of SPIE Proceedings, San Diego, USA, pages 1435-1446, May 2005. [PDF]
     Abstract: Surgical planning of liver tumor resections requires detailed three-dimensional (3D) understanding of the complex arrangement of vasculature, liver segments and tumors. Knowledge about location and sizes of liver segments is important for choosing an optimal surgical resection approach and predicting postoperative residual liver capacity. The aim of this work is to facilitate such surgical planning process by developing a robust method for portal vein tree segmentation. The work also investigates the impact of vessel segmentation on the approximation of liver segment volumes. For segment approximation, smaller portal vein branches are of importance. Small branches, however, are difficult to segment due to noise and partial volume effects. Our vessel segmentation is based on the original gray-values and on the result of a vessel enhancement filter. Validation of the developed portal vein segmentation method in computer generated phantoms shows that, compared to a conventional approach, more vessel branches can be segmented. Experiments with in vivo acquired liver CT data sets confirmed this result. The outcome of a Nearest Neighbor liver segment approximation method applied to phantom data demonstrates, that the proposed vessel segmentation approach translates into a more accurate segment partitioning. [bibtex-key = BeichelEtalSPIE2005]


180. Andrea Frosini, Simone Rinaldi, Elena Barcucci, and Attila Kuba. An efficient algorithm for reconstructing binary matrices from horizontal and vertical absorbed projections. In Proceedings of the Workshop on Discrete Tomography and its Applications, volume 20 of Electronic Notes in Discrete Mathematics, New York City, USA, pages 347-363, June 2005. [PDF] [doi:10.1016/j.endm.2005.05.073]
     Abstract: This paper studies the classical tomographical problem of the reconstruction of a binary matrix from projections in presence of absorption. In particular, we consider two projections along the horizontal and vertical directions and the mathematically interesting case of the absorption coefficient B0=(1+sqrt(5))/2. After proving some theoretical results on the switching components, we furnish a fast algorithm for solving the reconstruction problem from the horizontal and vertical absorbed projections. As a significative remark, we obtain also the solution of the related uniqueness problem. [bibtex-key = Frosini2005]


181. Zoltan Kiss, Lajos Rodek, Antal Nagy, Attila Kuba, and Marton Balasko. Reconstruction of pixel-based and geometric objects by discrete tomography. Simulation and physical experiments. In Proceedings of the Workshop on Discrete Tomography and its Applications, volume 20 of Electronic Notes in Discrete Mathematics, New York City, USA, pages 475-491, June 2005. [PDF] [doi:10.1016/j.endm.2005.05.080]
     Abstract: Tomography is an imaging technique to reconstruct cross-sections of objects from projection images in a non-destructive way. We have implemented two discrete tomographic (DT) reconstruction methods. The first method is general in the sense that the objects to be reconstructed are represented by pixels, and in this way the method can be used for reconstructing any shape. The second one is able to reconstruct objects consisting of cylinders and spheres made of homogeneous materials only. Simulation and physical experiments connected with non-destructive testing by X-rays were performed, and the results are presented. [bibtex-key = Kiss2005]


182. Sven Krimmel, Joachim Baumann, Zoltan Kiss, Attila Kuba, Antal Nagy, and Juergen Stephan. Discrete tomography for reconstruction from limited view angles in non-destructive testing. In Proceedings of the Workshop on Discrete Tomography and its Applications, volume 20 of Electronic Notes in Discrete Mathematics, pages 455-474, June 2005. [PDF] [doi:10.1016/j.endm.2005.05.078]
     Abstract: We show that discrete tomography (DT) is suitable to increase the possible inspection size of single material oblong objects compared to filtered back projection (FBP) in non-destructive testing (NDT) with 2D X-ray computed tomography (CT). For such objects which are in one dimension larger than the maximum detectable material thickness limited view angles occur and FBP is not suitable for reconstruction. For evaluation of the reconstruction performance a copper phantom (strong absorber) which exhibits typical problems for NDT was manufactured. The increase of the object size with DT reconstruction compared to FBP was estimated to be above 50%. [bibtex-key = Krimmel2005]


183. Attila Kuba, Lajos Rodek, Zoltan Kiss, Laszlo Rusko, Antal Nagy, and Marton Balasko. Discrete tomography in neutron radiography. In Proceedings of the International Topical Meeting on Neutron Radiography, volume 542 of Nuclear Instruments and Methods in Physics Research, Garching, Germany, pages 376-382, April 2005. [PDF] [doi:10.1016/j.nima.2005.01.164]
     Abstract: Discrete tomography (DT) is an imaging technique for reconstructing discrete images from their projections using the knowledge that the object to be reconstructed contains only a few homogeneous materials characterized by known discrete absorption values. One of the main reasons for applying DT is that we will hopefully require relatively few projections. Using discreteness and some a priori information (such as an approximate shape of the object) we can apply two DT methods in neutron imaging by reducing the problem to an optimization task. The first method is a special one because it is only suitable if the object is composed of cylinders and sphere shapes. The second method is a general one in the sense that it can be used for reconstructing objects of any shape. Software was developed and physical experiments performed in order to investigate the effects of several reconstruction parameters: the number of projections, noise levels, and complexity of the object to be reconstructed. We give a summary of the experimental results and make a comparison of the results obtained using a classical reconstruction technique (FBP). The programs we developed are available in our DT reconstruction program package DIRECT. [bibtex-key = Kuba2005a]


184. Attila Kuba, Laszlo Rusko, Zoltan Kiss, and Antal Nagy. Discrete Reconstruction Techniques. In Proceedings of the Workshop on Discrete Tomography and its Applications, volume 20 of Electronic Notes in Discrete Mathematics, New York City, USA, pages 385-398, June 2005. [PDF] [doi:10.1016/j.endm.2005.04.005]
     Abstract: DIscrete REConstruction Techniques (DIRECT) is a program system for generating test images, for computing their projections, for performing reconstructions, for comparing reconstructed images, and for visualizing the results. Our framework supplies a solid workflow to be performed during testing DT methods. A collection of DT reconstruction methods is planned to be implemented in this system to help further comparisions and software developments in DT. [bibtex-key = Kuba2005b]


185. Attila Kuba and Gerhard J. Woeginger. Two Remarks on Reconstructing Binary Vectors from Their Absorbed Projections. In Eric Andres, Guillaume Damiand, and Pascal Lienhardt, editors, Proceedings of the International Conference on Discrete Geometry for Computer Imagery, volume 3429 of Lecture Notes in Computer Science, Poitiers, France, pages 148-152, April 2005. Springer Verlag. [WWW] [doi:10.1007/b135490] [bibtex-key = Kuba2005c]


186. Antal Nagy, Attila Kuba, and Martin Samal. Reconstruction of factor structures using discrete tomography method. In Proceedings of the Workshop on Discrete Tomography and its Applications, volume 20 of Electronic Notes in Discrete Mathematics, New York City, USA, pages 519-534, June 2005. [PDF] [doi:10.1016/j.endm.2005.05.083]
     Abstract: Our aim was to estimate the volumes of homogeneous structures whose contrast / intensity was changing with time, using only few (here 4) projections of the structures. Each projection was recorded over a period of time and consisted of a sequence of images. The projections of the structures were first separated by factor analysis of the total projections. Then each structure was reconstructed individually from its factor projections using discrete tomography. Assumed application is the single-photon emission computed tomography (SPECT). [bibtex-key = NagyKuba:2005:Factor]


187. Laszlo G. Nyul, Judit Kanyo, Eors Mate, Geza Makay, Emese Balogh, Marta Fidrich, and Attila Kuba. Method for Automatically Segmenting the Spinal Cord and Canal from 3D CT Images. In A. Gagalowicz and W. Philips, editors, Proceedings of the International Conference on Computer Analysis of Images and Patterns, volume 3691 of Lecture Notes in Computer Science, Versailles, France, pages 456-463, September 2005. Springer Verlag. [PDF]
     Abstract: We present two approaches for automatically segmenting the spinal cord/canal from native CT images of the thorax region containing the spine. Different strategies are included to handle images where only part of the spinal column is visible. The algorithms require one seed point given on a slice located in the middle region of the spine, and the rest is automatic. The spatial extent of the spinal cord/canal is determined automatically. An extended region-growing technique is suggested for segmenting the spinal canal while active contours are applied if the spinal cord is to be segmented. Both methods work in 2D and use propagated information from neighboring slices. They are also very rapid in execution, that means an efficient, user-friendly workflow. The methods were evaluated by radiologists and were found to be useful (in reducing/ eliminating contouring labor and time) and met the accuracy and repeatability requirements for the particular task. [bibtex-key = Nyul:2005:MAS_CAIP]


188. Laszlo Rusko and Attila Kuba. Multi-resolution method for binary tomography. In Proceedings of the Workshop on Discrete Tomography and its Applications, volume 20 of Electronic Notes in Discrete Mathematics, New York City, USA, pages 299-311, July 2005. [PDF] [doi:10.1016/j.endm.2005.05.070]
     Abstract: Multi-resolution and region-growing strategies have been successfully used in several fields of image processing. In this paper we investigate how these two strategies can be applied for binary tomography. We included these strategies into a reconstruction method using simulated annealing and tested these new methods on different images. [bibtex-key = Rusko2005]


189. Attila Tanacs, Eors Mate, and Attila Kuba. Application of Automatic Image Registration in a Segmentation Framework for Pelvic CT Images. In Andre Gagalowicz and Wilfried Philips, editors, Proceedings of the International Conference on Computer Analysis of Images and Patterns, volume 3691 of Lecture Notes in Computer Science, Paris, France, pages 628-635, September 2005. [PDF] [doi:10.1007/11556121_77]
     Abstract: In radiation treatment (RT) planning, clinicians must trace the outline of a few critical structures on a large number of images. Using automated image segmentation could save tremendous time and effort. Segmentation of the organs near the pubic bone (prostate and bladder) is an important and challenging task: Some of the neighboring organs have similar density values in the CT images and the border between the different organs is hardly visible. In a segmentation framework, transforming a CT study to a common reference frame is used in two tasks: For statistical atlas (model) generation, and in the clinical application, establishing the voxel-to-voxel correspondence between the study and the model. In these cases precise alignment of all anatomical structures is not crucial, the focus is on proper alignment of the pubic bone area and fast execution. Our proposed method solves this by a new, two step process based on a voxel similarity-based registration algorithm. [bibtex-key = Tanacs:2005:CAIP]


190. Steffen Zopf and Attila Kuba. Reconstruction of measurable sets from two generalized projections. In Proceedings of the Workshop on Discrete Tomography and its Applications, volume 20 of Electronic Notes in Discrete Mathematics, New York City, USA, pages 47-66, June 2005. [PDF] [doi:10.1016/j.endm.2005.04.003]
     Abstract: The problem of reconstructing a measurable plane set from its two generalized projections is considered. It means that the projections contain also the effect of a known modification given in the whole plane. This is a more general case than that of a constant absorption within a given material. Via a suitable mapping, this generalized problem can be transformed into the solved case of the classical (non-absorbed and non-generalized) projections, giving a theorem about the characterization of unique, non-unique, and inconsistent projections analogous to Lorentz' theorem. The connection between uniqueness and the existence of so-called generalized switching components is discussed. [bibtex-key = Zopf2005]


191. Zoltan Kato. Reversible Jump Markov Chain Monte Carlo for Unsupervised MRF Color Image Segmentation. In Andreas Hoppe, Sarah Barman, and Tim Ellis, editors, Proceedings of the British Machine Vision Conference, volume 1, Kingston, UK, pages 37-46, September 2004. BMVA. [PDF] [PS]
     Abstract: Reversible jump Markov chain Monte Carlo (RJMCMC) is a recent method which makes it possible to construct reversible Markov chain samplers that jump between parameter subspaces of different dimensionality. In this paper, we propose a new RJMCMC sampler for multivariate Gaussian mixture identification and we apply it to color image segmentation. For this purpose, we consider a first order Markov random field (MRF) model where the singleton energies derive from a multivariate Gaussian distribution and second order potentials favor similar classes in neighboring pixels. The proposed algorithm finds the most likely number of classes, their associated model parameters and generates a segmentation of the image by classifying the pixels into these classes. The estimation is done according to the Maximum A Posteriori (MAP) criterion. Experimental results are promising, we have obtained accurate results on a variety of real color images. [bibtex-key = Kato2004a]


192. Kalman Palagyi, Juerg Tschirren, Eric A. Hoffman, and Milan Sonka. Assessment of Intrathoracic Airway Trees: Methods and In Vivo Validation. In Nicu Sebe, Michael S. Lew, and Thomas S. Huang, editors, Proceedings of the Computer Vision and Mathematical Methods in Medical and Biomedical Image Analysis: ECCV 2004 Workshops CVAMIA and MMBIA, volume 3117 of Lecture Notes in Computer Science, Prague, Czech Republic, pages 341-352, May 2004. Springer Verlag. [PDF]
     Abstract: A method for quantitative assessment of tree structures is reported allowing evaluation of airway tree morphology and its associated function. Our skeletonization and branch-point identification method provides a basis for tree quantification or tree matching, tree-branch diameter measurement in any orientation, and labeling individual branch segments. All main components of our method were specifically developed to deal with imaging artifacts typically present in volumetric medical image data. The proposed method has been tested in a computer phantom subjected to changes of its orientation as well as in repeatedly CT-scanned rigid and rubber plastic phantoms. In this paper, validation is reported in six in vivo scans of the human chest. [bibtex-key = PalagyiEtalECCV2004]


193. Ying Zhuge, Jayaram K. Udupa, and Laszlo G. Nyul. Multiple Sclerosis Lesion Quantification in MR Images by Using Vectorial Scale-based Relative Fuzzy Connectedness. In J. M. Fitzpatrick and M. Sonka, editors, Proceedings of Medical Imaging 2004: Image Processing, volume 5370 of SPIE Proceedings, San Diego, USA, pages 1764-1773, May 2004. [doi:10.1117/12.535655]
     Abstract: This paper presents a methodology for segmenting PD- and T2-weighted brain magnetic resonance (MR) images of multiple sclerosis (MS) patients into white matter (WM), gray matter (GM), cerebrospinal fluid (CSF), and MS lesions. For a given vectorial image (with PD- and T2-weighted components) to be segmented, we perform first intensity inhomogeneity correction and standardization prior to segmentation. Absolute fuzzy connectedness and certain morphological operations are utilized to generate the brain intracranial mask. The optimum thresholding method is applied to the product image (the image in which voxel values represent T2 value x PD value) to automatically recognize potential MS lesion sites. Then, the recently developed technique -- vectorial scale-based relative fuzzy connectedness -- is utilized to segment all voxels within the brain intracranial mask into WM, GM, CSF, and MS lesion regions. The number of segmented lesions and the volume of each lesion are finally output as well as the volume of other tissue regions. The method has been tested on 10 clinical brain MRI data sets of MS patients. An accuracy of better than 96% has been achieved. The preliminary results indicate that its performance is better than that of the k-nearest neighbors (kNN) method. [bibtex-key = Nyul:2004:MSL]


194. Peter Balazs, Emese Balogh, and Attila Kuba. A fast algorithm for reconstructing hv-convex 8-connected but not 4-connected discrete sets. In Ingela Nystrom, Gabriella Sanniti di Baja, and Stina Svensson, editors, Proceedings of the International Conference on Discrete Geometry for Computer Imagery, volume 2886 of Lecture Notes in Computer Science, Naples, Italy, pages 388-397, November 2003. Springer Verlag. [PDF]
     Abstract: One important class of discrete sets where the reconstruction from two given projections can be solved in polynomial time is the class of $hv$-convex 8-connected sets. The worst case complexity of the fastest algorithm known so far for solving the problem is of $O(mn\cdot\min\{ m^2,n^2\} )$. However, as it is shown, in the case of 8-connected but not 4-connected sets we can give an algorithm with worst case complexity of $O(mn\cdot\min\{ m,n\} )$ by identifying the so-called ${\cal S}_4$-components of the discrete set. Experimental results are also presented in order to investigate the average execution time of our algorithm. [bibtex-key = Balazs2003]


195. Xiaowen Ji, Zoltan Kato, and Zhiyong Huang. Non-Photorealistic Rendering and Content-Based Image Retrieval. In Proceedings of the Pacific Conference on Computer Graphics and Applications, Canmore, Canada, pages 153-162, October 2003. IEEE. [PDF]
     Abstract: In this paper, we will show how non-photorealistic rendering (NPR) can take a new role in content-based image retrieval (CBIR). The proposed CBIR method applies a novel image similarity measure: unlike traditional features like color, texture, or shape, our measure is based on a painted representation of the original image. This is produced by a stochastic paintbrush algorithm which simulates a painting process. We use the stroke parameters (color, size, orientation, and location) as features and similarity is measured by matching strokes of a pair of images. The advantage of our approach is that it provides information not only about the color content but also about the structural properties of an image without the segmentation of the image. Experimental results show that the CBIR method using paintbrush features has higher retrieval rate than traditional methods using color or texture features only. [bibtex-key = Ji-etal2003]


196. Zoltan Kato, Ting Chuen Pong, and Guo Qiang Song. Unsupervised segmentation of color textured images using a multi-layer MRF model. In Proceedings of the International Conference on Image Processing, volume I, Barcelona, Spain, pages 961-964, September 2003. IEEE. [PDF]
     Abstract: Herein, we propose a novel multilayer Markov random field (MRF) image segmentation model which aims at combining color and texture features: each feature is associated to a so called feature layer, where an MRF model is defined using only the corresponding feature. A special layer is assigned to the combined MRF model. This layer interacts with each feature layer and provides the segmentation based on the combination of different features. The model is quite generic and isn't restricted to a particular texture feature. Herein we will test the algorithm using Gabor and MRSAR texture features. Furthermore, the algorithm automatically estimates the number of classes at each layer (there can be different classes at different layers) and the associated model parameters. [bibtex-key = Kato-etal2003]


197. Kalman Palagyi, Juerg Tschirren, and Milan Sonka. Quantitative analysis of intrathoracic airway trees: methods and validation. In Chris Taylor and J. Alison Noble, editors, Proceedings of the International Conference on Information Processing in Medical Imaging (IPMI), volume 2732 of Lecture Notes in Computer Science, Ambleside, UK, pages 222-233, July 2003. Springer Verlag. [PDF]
     Abstract: A method for quantitative assessment of tree structures is reported allowing evaluation of airway or vascular tree morphology and its associated function. Our skeletonization and branch-point identification method provides a basis for tree quantification or tree matching, treebranch diameter measurement in any orientation, and labeling individual branch segments. All main components of our method were specifically developed to deal with imaging artifacts typically present in volumetric medical image data. The proposed method has been tested in 343 computer phantom instances subjected to changes of its orientation as well as in a repeatedly CT-scanned rubber plastic phantom width sub-voxel accuracy and high reproducibility. Application to 35 human in vivo trees yielded reliable and well-positioned centerlines and branch-points. [bibtex-key = PalagyiEtalIPMI2003]


198. Kalman Palagyi, Juerg Tschirren, and Milan Sonka. Quantitative analysis of three-dimensional tubular tree structures. In J. M. Fitzpatrick and M. Sonka, editors, Proceedings of Medical Imaging 2004: Image Processing, volume 5032 of SPIE Proceedings, San Diego, USA, pages 277-287, May 2003. [PDF]
     Abstract: Quantitative assessment of tree structures is very important for evaluation of airway or vascular tree morphology and its associated function. Our skeletonization and branch-point identification method provides a basis for tree quantification or tree matching, tree-branch diameter measurement in any orientation, and labeling individual branch segments. All main components of our method were specifically developed to deal with imaging artifacts typically present in volumetric medical image data. The proposed method has been tested in a computer phantom subjected to changes of its orientation as well as in a repeatedly CT-scanned rigid plastic phantom. In all cases, our method produced reliable and well positioned centerlines and branch-points. [bibtex-key = PalagyiEtalSPIE2003]


199. Emese Balogh, Erich Sorantin, Laszlo G. Nyul, Kalman Palagyi, Attila Kuba, Georg Werkgartner, and Ekke Spuller. Virtual Dissection of the Colon -- Technique and First Experiments with Artificial and Cadaveric Phantoms. In S. K. Mun, editor, Proceedings of Medical Imaging 2002: Visualization, Image-Guided Procedures, and Display, volume 4681 of SPIE Proceedings, San Diego, USA, pages 713-721, May 2002. SPIE. [doi:10.1117/12.466982]
     Abstract: Virtual dissection refers to a display technique for polyp detection, where the colon is digitally straightened and then flattened using multirow detector Computed Tomograph (CT) images. As compared to virtual colonoscopy where polyps may be hidden from view behind the folds, the unravelled colon is more suitable for polyp detection, because the entire inner surface of the colon is displayed in a single view. The method was tested both on artificial and cadaveric phantoms. All polyps could be recognized on both phantoms. This technique for virtual dissection requires only a minimum of operator interaction. [bibtex-key = Balogh:2002:VDC]


200. Gabor Fichtinger, Axel Krieger, Robert C. Susil, Attila Tanacs, Louis L. Whitcomb, and Ergin Atalar. Transrectal prostate biopsy inside closed MRI scanner with remote actuation, under real-time image guidance. In Takeyoshi Dohi and Ron Kikins, editors, Proceedings of the International Conference on Medical Image Computing & Computer Assisted Intervention (MICCAI), volume 2488 (1) of Lecture Notes in Computer Science, Tokyo, Japan, pages 91-98, September 2002. [PDF]
     Abstract: We present the proof-of-concept prototype of a prostate biopsy robot to be used inside a conventional high-field MRI scanner. A three degree-of-freedom (DOF) mechanical device translates and rotates inside the rectum and enters a needle into the body, and steers the needle to a target point pre-selected by the user. The device is guided by real-time images from the scanner. Networked computers process the medical images and enable the clinician to control the motion of the mechanical device that is operated remotely from outside the imager. The system is also applicable to localized prostate therapy and also demonstrates potential in other intra-cavitary procedures. [bibtex-key = Fichtinger:2002:MICCAI]


201. Zoltan Kato, Xiaowen Ji, Tamas Sziranyi, Zoltan Toth, and Laszlo Czuni. Content-Based Image Retrieval Using Stochastic Paintbrush Transformation. In Proceedings of the International Conference on Image Processing, volume 1, New York, USA, pages 944-947, September 2002. IEEE. [PDF] [PS]
     Abstract: We propose a new content based image retrieval method. The novelty of our approach lies in the applied image similarity measure: unlike traditional features, such as color, texture or shape, our measure is based on a painted representation of the original image. We use paintbrush stroke parameters as features. These strokes are produced by a stochastic paintbrush algorithm which simulates a painting process. Stroke parameters include color, orientation and location. Therefore, it provides information not only about the color content but also about the structural properties of an image. Experimental results on a database of more than 500 images show that the CBIR method using paintbrush features has a higher retrieval rate than methods using color features only. [bibtex-key = Kato-etal2002a]


202. Zoltan Kato, Ting Chuen Pong, and Guo Qiang Song. Multicue MRF Image Segmentation: Combining Texture and Color. In Proceedings of the International Conference on Pattern Recognition, volume 1, Quebec, Canada, pages 660-663, August 2002. IAPR, IEEE. [PDF] [PS]
     Abstract: We propose a new Markov random field (MRF) image segmentation model which aims at combining color and texture features. The model has a multilayer structure: each feature has its own layer, called feature layer, where an MRF model is defined using only the corresponding feature. A special layer is assigned to the combined MRF model. This layer interacts with each feature layer and provides the segmentation based on the combination of different features. The uniqueness of our algorithm is that it provides both color only and texture only segmentations as well as a segmentation based on combined color and texture features. The number of classes on feature layers is given by the user, but it is estimated on the combined layer. [bibtex-key = Kato-etal2002]


203. Laszlo G. Nyul and Jayaram K. Udupa. A Protocol-independent Brain MRI Segmentation Method. In M. Sonka and J. M. Fitzpatrick, editors, Proceedings of Medical Imaging 2002: Image Processing, volume 4684 of SPIE Proceedings, San Diego, USA, pages 1588-1599, May 2002. [PDF]
     Abstract: We present a segmentation method that combines the robust, accurate, and efficient techniques of fuzzy connectedness with standardized MRI intensities and fast algorithms. The result is a general segmentation framework that more efficiently utilizes the user input (for recognition) and the power of computer (for delineation). This same method has been applied to segment brain tissues from a variety of MRI protocols. Images were corrected for inhomogeneity and standardized to yield tissue-specific intensity values. All parameters for the fuzzy affinity relations were fixed for a specific input protocol. Scale-based fuzzy affinity was used to better capture fine structures. Brain tissues were segmented as 3D fuzzy-connected objects by using relative fuzzy connectedness. The user can specify seed points in about a minute and tracking the 3D fuzzy-connected objects takes about 20 seconds per object. All other computations were performed before any user interaction took place. Segmentation of brain tissues as 3D fuzzy-connected objects from MRI data is feasible at interactive speeds. Utilizing the robust fuzzy connectedness principles and fast algorithms, it is possible to interactively select fuzzy affinity, seed point, and threshold parameters and perform efficient, precise, and accurate segmentations. [bibtex-key = Nyul:2002:PBM]


204. Juerg Tschirren, Kalman Palagyi, Joseph M. Reinhardt, Eric A. Hoffman, and Milan Sonka. Segmentation, skeletonization, and branchpoint matching - A fully automated quantitative evaluation of human intrathoratic airway trees. In Takeyoshi Dohi and Ron Kikinis, editors, Proceedings of the International Conference on Medical Image Computing & Computer Assisted Intervention (MICCAI), volume 2489 of Lecture Notes in Computer Science, Tokyo, Japan, pages 12-19, September 2002. Springer Verlag. [PDF]
     Abstract: Modern multislice X-ray CT scanners provide high-resolution volumetric image data containing a wealth of structural and functional information. The size of the volumes makes it more and more di?cult for human observers to visually evaluate their contents. Similar to other areas of medical image analysis, highly automated extraction and quantitative assessment of volumetric data is increasingly important in pulmonary physiology, diagnosis, and treatment. We present a method for a fully automated segmentation of a human airway tree, its skeletonization, identi?cation of airway branches and branchpoints, as well as a method for matching the airway trees, branches, and branchpoints for the same subject over time and across subjects. The validation of our method shows a high correlation between the automatically obtained results and reference data provided by human observers. [bibtex-key = TschirrenEtalMICCAI2002]


205. Gabor Fichtinger, Ken Masamune, Alexandru Patriciu, Attila Tanacs, James H. Anderson, Theodore L. DeWeese, Russel H. Taylor, and Dan Stoianovici. Robotically Assisted Percutaneous Local Therapy and Biopsy. In Proceedings of the IEEE International Conference on Advanced Robotics, Budapest, Hungary, pages 133-151, August 2001. IEEE. [PDF]
     Abstract: We present the concept and prototype of an image-guided robotic system for accurate and consistent placement of percutaneous needles in soft tissue targets under CT-guidance. The system is a promising embodiment of the Surgical CAD/CAM paradigm and as such, easily adaptable to other image guidance modalities, like X-ray fluoroscopy. We also report the first results of pre-clinical experiments on phantoms. [bibtex-key = Fichtinger:2001:ICAR]


206. Gabor Fichtinger, Sheng Xu, Attila Tanacs, Kieran Murphy, Lee Myers, and Jeffery Williams. Approximate Volumetric Reconstruction from Projected Images. In Wiro J. Niessen and Max A. Viergever, editors, Proceedings of the International Conference on Medical Image Computing & Computer Assisted Intervention (MICCAI), volume 2208 of Lecture Notes in Computer Science, Utrecht, The Netherlands, pages 1376-1378, October 2001. Springer Verlag. [PDF]
     Abstract: A significant problem in planning of volumetrically prescribed localized treatments is the mathematical impossibility to determine the exact three dimensional shape and volume of a target object from its projected images. Reconstruction accuracy also varies with viewing angle, depending on the convexity and aspect ratios of the target object. In response to this problem, we are developing a robust and efficient technique for approximate volumetric reconstruction, which (A) uses no prior information of the shape and volume of the target, (B) does not require exact silhouettes, (C) accepts arbitrary number of images, (D) produces solid object and measure of its volume, (E) provides confidence measure of the reconstruction and drawing of silhouettes, (F) is robust, fast and easy to implement. Preliminary tests suggest that fairly convex objects can be reconstructed from four views, and typically six views with table rotation allow us to reconstruct fine details as small as 1 mm. The method is applicable for any X-ray guided volumetric treatment. Pilot applications will be planning of radiosurgery of arterioveneous malformations (AVMs) and radiofrequency ablation of soft tissue lesions. [bibtex-key = Fichtinger:2001:MICCAI]


207. Zoltan Kato and Ting Chuen Pong. A Markov Random Field Image Segmentation Model Using Combined Color and Texture Features. In Wladyslaw Skarbek, editor, Proceedings of the International Conference on Computer Analysis of Images and Patterns, volume 2124 of Lecture Notes in Computer Science, Warsaw, Poland, pages 547-554, September 2001. Springer Verlag. [PDF] Keyword(s): image segmentation, Markov random field model.
     Abstract: In this paper, we propose a Markov random field (MRF) image segmentation model which aims at combining color and texture features. The theoretical framework relies on Bayesian estimation associated with combinatorial optimization (Simulated Annealing). The segmentation is obtained by classifying the pixels into different pixel classes. These classes are represented by multi-variate Gaussian distributions. Thus, the only hypothesis about the nature of the features is that an additive white noise model is suitable to describe the feature values belonging to a given class. Herein, we use the perceptually uniform CIE-L*u*v* color values as color features and a set of Gabor filters as texture features. We provide experimental results that illustrate the performance of our method on both synthetic and natural color images. Due to the local nature of our MRF model, the algorithm can be highly parallelized. [bibtex-key = Kato-Pong2001]


208. Attila Kuba and Antal Nagy. Reconstruction of hv-convex Binary Matrices from Their Absorbed Projections. In Proceedings of the International Workshop on Combinatorial Image Analysis, volume 46 of Electronic Notes in Theoretical Computer Science, Philadephia, Pennsylvani, USA, pages 371-380, August 2001. [PDF] [doi:10.1016/S1571-0661(04)80998-8]
     Abstract: The reconstruction of hv-convex binary matrices from their absorbed projections is considered. Although this problem is NP-complete if the non-absorbed row and column sums are available, it is proved that such a reconstruction problem can be solved in polynomial time from absorbed projections when the absorption is represented by B = (1 + sqrt(5))/2. Also a reconstruction algorithm is given to determine the whole structure of hv-convex binary matrices from such projections. [bibtex-key = Kuba2001]


209. Laszlo G. Nyul, Jayaram K. Udupa, and Punam K. Saha. Task-specific Comparison of 3-D Image Registration methods. In M. Sonka and K. M. Hanson, editors, Proceedings of Medical Imaging 2001: Image Processing, volume 4322 of SPIE Proceedings, San Diego, USA, pages 1588-1598, July 2001. [doi:10.1117/12.431044]
     Abstract: We present a new class of approaches for rigid-body registration and their evaluation in studying Multiple Sclerosis via multi protocol MRI. Two pairs of rigid-body registration algorithms were implemented, using cross- correlation and mutual information, operating on original gray-level images and on the intermediate images resulting from our new scale-based method. In the scale image, every voxel has the local scale value assigned to it, defined as the radius of the largest sphere centered at the voxel with homogeneous intensities. 3D data of the head were acquired from 10 MS patients using 6 MRI protocols. Images in some of the protocols have been acquired in registration. The co-registered pairs were used as ground truth. Accuracy and consistency of the 4 registration methods were measured within and between protocols for known amounts of misregistrations. Our analysis indicates that there is no best method. For medium and large misregistration, methods using mutual information, for small misregistration, and for the consistency tests, correlation methods using the original gray-level images give the best results. We have previously demonstrated the use of local scale information in fuzzy connectedness segmentation and image filtering. Scale may also have considerable potential for image registration as suggested by this work. [bibtex-key = Nyul:2001:TC3]


210. Kalman Palagyi, Erich Sorantin, Emese Balogh, Attila Kuba, Csongor Halmai, Balazs Erdohelyi, and Klaus Hausegger. A sequential 3D thinning algorithm and its medical applications. In Michael F. Insana and Richard M. Leahy, editors, Proceedings of the International Conference on Information Processing in Medical Imaging (IPMI), volume 2082 of Lecture Notes in Computer Science, Davis, CA, USA, pages 409-415, June 2001. Springer Verlag. [PDF]
     Abstract: Skeleton is a frequently applied shape feature to represent the general form of an object. Thinning is an iterative object reduction technique for producing a reasonable approximation to the skeleton in a topology preserving way. This paper describes a sequential 3D thinning algorithm for extracting medial lines of objects in (26,6) pictures. Our algorithm has been successfully applied in medical image analysis. Three of the emerged applications (analysing airways, blood vessels, and colons) are also presented. [bibtex-key = PalagyiEtalIPMI2001]


211. Attila Kuba and Maurice Nivat. Reconstruction of Discrete Sets with Absorption. In Gunilla Borgefors, Ingela Nystrom, and Gabriella Sanniti di Baja, editors, Proceedings of the International Conference on Discrete Geometry for Computer Imagery, volume 1953 of Lecture Notes in Computer Science, Uppsala, Sweden, pages 137, 2000. Springer Verlag. [WWW] [PDF]
     Abstract: A generalization of a classical discrete tomography problem is considered: Reconstruct binary matrices from their absorbed row and columns sums, i.e., when some known absorption is supposed. It is mathematically interesting when the absorbed projection of a matrix element is the same as the absorbed projection of the next two consecutive elements together. We show that, in this special case, the non-uniquely determined matrices contain a certain configuration of 0s and 1s, called alternatively corner-connected components. Furthermore, such matrices can be transformed into each other by switchings the 0s and 1s of these components. [bibtex-key = Kuba2000]


212. Laszlo G. Nyul, Alexandre X. Falcao, and Jayaram K. Udupa. Fuzzy-connected 3D Image Segmentation at Interactive Speeds. In K. M. Hanson, editor, Proceedings of Medical Imaging 2000: Image Processing, volume 3979 of SPIE Proceedings, San Diego, USA, pages 212-223, June 2000. [PDF]
     Abstract: Image segmentation techniques using fuzzy connectedness principles have shown their effectiveness in segmenting a variety of objects in several large applications in recent years. However, one problem with these algorithms has been their excessive computational requirements. In an attempt to substantially speed them up, in the present paper, we study systematically a host of 18 algorithms under two categories -- label correcting and label setting. Extensive testing of these algorithms on a variety of 3D medical images taken from large ongoing applications demonstrates that a 20-360 fold improvement over current speeds is achievable with a combination of algorithms and fast modern PCs. The reliable recognition (assisted by human operators) and the accurate, efficient, and sophisticated delineation (automatically performed by the computer) can be effectively incorporated into a single interactive process. If images having intensities with tissue specific meaning (such as CT or standardized MR images) are utilized, all parameters for the segmentation method can be fixed once for all, all intermediate data can be computed before the user interaction is needed, and the user can be provided with more information at the time of interaction. [bibtex-key = Nyul:2000:FC3]


213. Laszlo G. Nyul and Jayaram K. Udupa. Standardizing the MR Image Intensity Scales: Making MR Intensities Have Tissue-specific Meaning. In S. K. Mun, editor, Proceedings of Medical Imaging 2000: Image Display and Visualization, volume 3976 of SPIE Proceedings, San Diego, USA, pages 496-504, April 2000. [doi:10.1117/12.383076]
     Abstract: One of the major drawbacks of Magnetic Resonance Imaging (MRI) has been the lack of a standard and quantifiable interpretation of image intensities. Unlike in other modalities such as x-ray computerized tomography, MR images taken for the same patient on the same scanner at different times may appear different from each other due to a variety of scanner-dependent variations, and therefore, the absolute intensity values do not have a fixed meaning. We have devised a two-step method wherein all images can be transformed in such a way that for the same protocol and body region, in the transformed images similar intensities will have similar tissue meaning. Standardized images can be displayed with fixed windows without the need of per case adjustment. More importantly, extraction of quantitative information with fixed windows without the need of per case adjustment. More importantly, extraction of quantitative information about healthy organs or about abnormalities can be considerably simplified. This paper introduces and compares new variants of this standardizing method that can help to overcome some of the problems with the original method. [bibtex-key = Nyul:2000:SMI]


214. Kalman Palagyi. A 3D 3-subiteration thinning algorithm for medial surfaces. In Gunilla Borgefors, Ingela Nystrom, and Gabriella Sanniti di Baja, editors, Proceedings of the International Conference on Discrete Geometry for Computer Imagery, volume 1953 of Lecture Notes in Computer Science, Uppsala, Sweden, pages 406-417, December 2000. Springer Verlag. [PDF]
     Abstract: Thinning on a binary picture is an iterative layer by layer erosion to extract a reasonable approximation to its skeleton. This paper presents an efficient 3D parallel thinning algorithm which produces medial surfaces. Three-subiteration directional strategy is proposed: each iteration step is composed of three parallel subiterations according to the three deletion directions. The algorithm makes easy implementation possible, since deletable points are given by matching templates containing twentyeight elements. The topological correctness of the algorithm for (26,6) binary pictures is proved. [bibtex-key = PalagyiDGCI2000]


215. Attila Tanacs, Gabor Czedli, Kalman Palagyi, and Attila Kuba. Point-based registration assuming affine motion. In Gerald Sommer and Yehoshua Y. Zeevi, editors, Proceedings of the International Workshop on Algebraic Frames for the Perception-Action Cycle (AFPAC), volume 1888 of Lecture Notes in Computer Science, Kiel, Germany, pages 329-338, September 2000. Springer Verlag. [PDF] [doi:10.1007/10722492_26]
     Abstract: Registration is a fundamental task in image processing. Its purpose is to find a geometrical transformation that relates the points of an image to their corresponding points of another image. The determination of the optimal transformation depends on the types of variations between the images. In this paper we propose a robust method based on two sets of points representing the images. One--to--one correspondence is assumed between these two sets. Our approach finds global affine transformation between the sets of points and can be used in any arbitrary dimension $k\ge 1$. A sufficient existence condition for a unique solution is given and proven. Our method can be used to solve various registration problems emerged in numerous fields, including medical image processing, remotely sensed data processing, and computer vision. [bibtex-key = Tanacs:2000:AFPAC]


216. Jayaram K. Udupa, Laszlo G. Nyul, Yulin Ge, and Robert I. Grossman. Multiprotocol MR Image Segmentation in Multiple Sclerosis: Experience With Over 1000 Studies. In K. M. Hanson, editor, Proceedings of Medical Imaging 2000: Image Processing, volume 3979 of SPIE Proceedings, San Diego, USA, pages 1017-1027, June 2000. [doi:10.1117/12.387606]
     Abstract: Multiple Sclerosis (MS) is an acquired disease of the central nervous system. Subjective cognitive and ambulatory test scores on a scale called EDSS are currently utilized to assess the disease severity. Various MRI protocols are being investigated to study the disease based on how it manifests itself in the images. In an attempt to eventually replace EDSS by an objective measure to assess the natural course of the disease and its response to therapy, we have developed image segmentation methods based on fuzzy connectedness to quantify various objects in multiprotocol MRI. These include the macroscopic objects such as lesions, the gray matter (GM), white matter (WM), cerebrospinal fluid (CSF), and brain parenchyma as well as the microscopic aspects of the diseased WM. Over 1000 studies have been processed to date. By far the strongest correlations with the clinical measures were demonstrated by the Magnetization Transfer Ratio (MTR) histogram parameters obtained for the various segmented tissue regions emphasizing the importance of considering the microscopic/diffused nature of the disease in the individual tissue regions. Brain parenchymal volume also demonstrated a strong correlation with the clinical measures indicating that brain atrophy is an important indicator of the disease. Fuzzy connectedness is a viable segmentation method for studying MS. [bibtex-key = Udupa:2000:MMI]


217. Endre Katona and Gyorgy Hudra. An Interpretation System for Cadastral Maps. In Proceedings of the International Conference on Image Analysis and Processing, Venice, Italy, pages 792-797, September 1999. IEEE. [PDF]
     Abstract: To create a spatial database for some GIS application, it is a big challenge to recognize automatically all the simple and complex map objects on scanned maps. This paper presents a robust map interpretation system developed to process Hungarian land register maps (cadastral maps). Processing starts with a raster-to-vector conversion generating a raw vector image from the scanned map. All recognition steps are performed on this raw vector image: segmentation, recognition of separated and not separated symbols, recognition of more complex objects (buildings and parcels). Finally, drawing quality is enhanced utilizing the recognition results. Interpretation is supported by a special data structure - called DG - ensuring dynamic description of hierarchical structures of drawing objects. This data structure is an essential part of our concept, therefore it is discussed in the paper. [bibtex-key = Katona1999]


218. Attila Kuba. Reconstruction in Different Classes of 2D Discrete Sets. In Gilles Bertrand, Michel Couprie, and Laurent Perroton, editors, Proceedings of the International Conference on Discrete Geometry for Computer Imagery, volume 1568 of Lecture Notes in Computer Science, Marne-la-Vallee, France, pages 153, March 1999. Springer Verlag. [WWW] [PDF]
     Abstract: The problem of reconstruction of two-dimensional discrete sets from their two projections is considered in different classes. The reconstruction algorithms and complexity results are summarized in the case of hv-convex sets, hv-convex polyominoes, hv-convex 8-connected sets, and directed h-convex sets. We show that the reconstruction algorithms used in the class of hv-convex 4-connected sets (polyominoes) can be used, with small modifications, for reconstructing hv-convex 8-connected sets. Finally, it is shown that the directed h-convex sets are uniquely reconstructible with respect to the row and column sum vectors. [bibtex-key = Kuba1999a]


219. Laszlo G. Nyul and Jayaram K. Udupa. An Approach to Standardizing the MR Image Intensity Scale. In S. K. Mun and Y. Kim, editors, Proceedings of Medical Imaging 1999: Image Display, volume 3658 of SPIE Proceedings, San Diego, USA, pages 595-603, May 1999. [doi:10.1117/12.349472]
     Abstract: Despite the many advantages of MR images, they lack a standard image intensity scale. MR image intensity ranges and the meaning of intensity values vary even for the same protocol (P) and the same body region (D). This causes many difficulties in image display and analysis. We propose a two-step method for standardizing the intensity scale in such a way that for the same P and D, similar intensities will have similar meanings. In the first step, the parameters of the standardizing transformation are 'learned' from an image set. In the second step, for each MR study, these parameters are used to map their histogram into the standardized histogram. The method was tested quantitatively on 90 whole brain FSE T2, PD and T1 studies of MS patients and qualitatively on several other SE PD, T2 and SPGR studies of the grain and foot. Measurements using mean squared difference showed that the standardized image intensities have statistically significantly more consistent range and meaning than the originals. Fixed windows can be established for standardized imags and used for display without the need of per case adjustment. Preliminary results also indicate that the method facilitates improving the degree of automation of image segmentation. [bibtex-key = Nyul:1999:ASM]


220. Laszlo G. Nyul and Jayaram K. Udupa. New Variants of a Method of MRI Scale Normalization. In A. Kuba, M. Samal, and A. Todd-Pokropek, editors, Proceedings of the International Conference on Information Processing in Medical Imaging (IPMI), volume 1613 of Lecture Notes in Computer Science, Visegrad, Hungary, pages 490-495, jun--jul 1999. Springer Verlag. [PDF]
     Abstract: One of the major drawbacks of Magnetic Resonance Imaging (MRI) has been the lack of a standard and quantifiable interpretation of image intensities. This causes many difficulties in image display and analysis. We have devised a two-step method wherein all images can be transformed in such a way that for the same protocol and body region, in the transformed images similar intensities will have similar tissue mean- ing. Normalized images can be displayed with fixed windows without the need of per case adjustment. More importantly, extraction of quantitative information about healthy organs or about abnormities, such as tumors, can considerably be simplified. This paper introduces and compares new variants of this normalization method that can help to overcome some of the problems with the original method. [bibtex-key = Nyul:1999:NVM_IPMI]


221. Kalman Palagyi and Attila Kuba. Directional 3D thinning using 8 subiterations. In Gilles Bertrand, Michel Couprie, and Laurent Perroton, editors, Proceedings of the International Conference on Discrete Geometry for Computer Imagery, volume 1568 of Lecture Notes in Computer Science, Marne-la-Vallee, France, pages 325-336, March 1999. Springer Verlag. [PDF]
     Abstract: Thinning of a binary object is an iterative layer by layer erosion to extract an approximation to its skeleton. In order to provide topology preservation, different thinning techniques have been proposed. One of them is the directional (or border sequential) approach in which each iteration step is subdivided into subiterations where only border points of certain kind are deleted in each subiteration. There are six kinds of border points in 3D images, therefore, 6-subiteration parallel thinning algorithms were generally proposed. In this paper, we present two 8-subiteration algorithms for extracting "surface skeletons" and "curve skeletons", respectively. Both algorithms work in cubic grid for (26,6) images. Deletable points are given by templates that makes easy implementation possible. [bibtex-key = PalagyiKubaDGCI1999]


222. Zoltan Kato, Ting Chuen Pong, and John Chung Mong Lee. Motion Compensated Color Video Classification Using Markov Random Fields. In Roland Chin and Ting Chuen Pong, editors, Proceedings of the Asian Conference on Computer Vision, volume 1351 of Lecture Notes in Computer Science, Hong Kong, China, pages 738-745, January 1998. Springer Verlag. [PDF] [PS]
     Abstract: This paper deals with the classification of color video sequences using Markov Random Fields (MRF) taking into account motion information. The theoretical framework relies on Bayesian estimation associated with MRF modelization and combinatorial optimization (Simulated Annealing). In the MRF model, we use the CIE-luv color metric because it is close to human perception when computing color differences. In addition, intensity and chroma information is separated in this space. The sequence is regarded as a stack of frames and both intra- and inter-frame cliques are defined in the label field. Without motion compensation, an inter-frame clique would contain the corresponding pixel in the previous and next frame. In the motion compensated model, we add a displacement field and it is taken into account in inter-frame interactions. The displacement field is also a MRF but there are no inter-frame cliques. The Maximum A Posteriori (MAP) estimate of the label and displacement field is obtained through Simulated Annealing. Parameter estimation is also considered in the paper and results are shown on color video sequences using both the simple and motion compensated models. [bibtex-key = Kato-etal98]


223. Tamas Sziranyi, Josiane Zerubia, David Geldreich, and Zoltan Kato. Cellular Neural Network in Markov Random Field Image Segmentation. In Proceedings of the International Workshop on Cellular Neural Networks and their Applications, Seville, Spain, pages 139-144, June 1996. IEEE.
     Abstract: Statistical approaches to early vision processes need a huge amount of computing power. These algorithms can usually be implemented on parallel computing structures. CNN is a fast parallel processor array for image processing. However, CNN is basically a deterministic analog circuit. We use the CNN-UM architecture for statistical image segmentation. With a single random in-put signal, we were able to implement a (pseudo) random field generator using one layer (one memory/cell) of the CNN. The whole algorithm needs 8 memories/cell. We can introduce this pseudo-stochastic segmentation process in the CNN structure. Considering the simple structure of the analog VLSI design, we use simple arithmetic functions (addition, multiplication) and very simple nonlinear output functions (step, jigsaw). With this architecture, a real VLSI CNN chip can execute a pseudo-stochastic relaxation algorithm of about 100 iterations in about 1 msec. In the Markov random field (MRF) theory, one important problem is parameter estimation. The random segmentation process must be preceded by the estimation of the gray-level distribution of the different classes on small image segments. This process is basically supervised. Usually the histograms of noisy images can be modelled as simple Gaussian distributions. This approach cannot be held in a CNN structure, since there should be as many additional layers as the number of classes. We should follow another way. We have developed a pixel-level distribution model. [bibtex-key = Sziranyi-etal96]


224. Zoltan Kato, Mark Berthod, Josiane Zerubia, and Wojciech Pieczynski. Unsupervised Adaptive Image Segmentation. In Proceedings of the International Conference on Acoustics, Speech and Signal Processing, volume 4, Detroit, Michigan, USA, pages 2399-2402, May 1995. IEEE. [PDF]
     Abstract: This paper deals with the problem of unsupervised Bayesian segmentation of images modeled by Markov random fields (MRF). If the model parameters are known then we have various methods to solve the segmentation problem (simulated annealing, ICM, etc...). However, when they are not known, the problem becomes more difficult. One has to estimate the hidden label field parameters from the available image only. Our approach consists of a recent iterative method of estimation, called iterative conditional estimation (ICE), applied to a monogrid Markovian image segmentation model. The method has been tested on synthetic and real satellite images. [bibtex-key = Kato-etal95a]


225. Zoltan Kato, Josiane Zerubia, and Mark Berthod. Unsupervised Parallel Image Classification Using a Hierarchical Markovian Model. In Proceedings of the International Conference on Computer Vision, Cambridge, MA, USA, pages 169-174, June 1995. IEEE. [PDF] [PS]
     Abstract: This paper deals with the problem of unsupervised classification of images modeled by Markov Random Fields (MRF). If the model parameters are known then we have various methods to solve the segmentation problem (simulated annealing, ICM, etc\ldots ). However, when they are not known, the problem becomes more difficult. One has to estimate the hidden label field parameters from the only observable image. Our approach consists of extending a recent iterative method of estimation, called Iterative Conditional Estimation (ICE) to a hierarchical markovian model. The idea resembles the Estimation-Maximization (EM) algorithm as we recursively look at the Maximum a Posteriori (MAP) estimate of the label field given the estimated parameters then we look at the Maximum Likelihood (ML) estimate of the parameters given a tentative labeling obtained at the previous step. We propose unsupervised image classification algorithms using a hierarchical model. The only parameter supposed to be known is the number of regions, all the other parameters are estimated. The presented algorithms have been implemented on a Connection Machine CM200. Comparative tests have been done on noisy synthetic and real images (remote sensing). [bibtex-key = Kato-etal95b]


226. Endre Katona, Kalman Palagyi, and Nandor Toth. Signature verification using neural nets. In Proceedings of the Scandinavian Conference on Image Analysis, Uppsala, Sweden, pages 1115-1122, June 1995. [PDF]
     Abstract: It is a hard problem to decide whether two signatures - given as scanned binary images - are written by the same person or not. The present paper gives a complex strategy to solve this problem, including thinning of the scanned image, raster-to-vector conversion, extracting features from vectorgraph representation and a decision process using a back propagation neural network model. Some aspects of neural fine tuning and learning are discussed. Features and experiences of the software implementation are mentioned. [bibtex-key = Katona1995]


227. Josiane Zerubia, Zoltan Kato, and Mark Berthod. Multi-Temperature Annealing: A New Approach for the Energy-Minimization of Hierarchical Markov Random Field Models. In Proceedings of the International Conference on Pattern Recognition, volume 1, Jerusalem, Israel, pages 520-522, October 1994. IAPR. [PS]
     Abstract: As it is well known, optimization of the energy function of Markov Random Fields is very expensive. Hierarchical models have usually much more communication per pixel than monogrid ones. This is why classical annealing schemes are too slow, even on a parallel machine, to minimize the energy associated with such a model. However, taking benefit of the pyramidal structure of the model, we can define a new annealing scheme: the Multi-Temperature Annealing (MTA), which consists of associating higher temperatures to coarser levels, in order to be less sensitive to local minima at coarser grids. The convergence to the global optimum is proved by a generalisation of the annealing theorem of Geman and Geman. We have applied the algorithm to image classification and tested it on synthetic and real images. [bibtex-key = Zerubia-etal94]


228. Zoltan Kato, Mark Berthod, and Josiane Zerubia. A Hierarchical Markov Random Field Model for Image Classification. In Proceedings of the International Workshop on Image and Multidimensional Digital Signal Processing, Cannes, France, September 1993. IEEE. [PS]
     Abstract: In this paper, we propose a hierarchical Markov Random Field (MRF) model. This model is based on a classical multiscale model, which consists of a label pyramid and a whole observation field. The parameters of the coarse grid can be derived by simple computation from the finest grid. In the label pyramid, we have introduced a new local interaction between two neighbor grids. This model gives a relaxation algorithm with a new annealingscheme: The Multi-Temperature Annealing (MTA) scheme, which consists of associating higher temperatures to higher levels, thus beeing less sensitive to local minima at coarser grids. The model was tested on different synthetic and real images. The algorithm was implemented on a Connection MachineCM200. [bibtex-key = Kato-etal93d]


229. Zoltan Kato, Mark Berthod, and Josiane Zerubia. Multiscale Markov Random Field Models for Parallel Image Classification. In Proceedings of the International Conference on Computer Vision, Berlin, Germany, pages 253-257, May 1993. IEEE. [PS]
     Abstract: In this paper, we are interested in multiscale Markov Random Field (MRF) models. It is well known that multigrid methods can improve significantly the convergence rate and the quality of the final results of iterative relaxation techniques. Herein, we propose a new hierarchical model, which consists of a label pyramid and a whole observation field. The parameters of the coarse grid can be derived by simple computation from the finest grid. In the label pyramid, we have introduced a new local interaction between two neighbor grids. This model gives a relaxation algorithm which can be run in parallel on the entire pyramid. On the other hand, the new model allows to propagate local interactions more efficiently giving estimates closer to the global optimum for deterministic as well as for stochastic relaxation schemes. It can also be seen as a way to incorporate cliques with far apart sites for a reasonable price. [bibtex-key = Kato-etal93a]


230. Zoltan Kato, Mark Berthod, and Josiane Zerubia. Parallel Image Classification using Multiscale Markov Random Fields. In Proceedings of the International Conference on Acoustics, Speech and Signal Processing, volume 5, Minneapolis, USA, pages 137-140, April 1993. IEEE. [PS]
     Abstract: In this paper, we are interested in massively parallel multiscale relaxation algorithms applied to image classification. First, we present a classical multiscale model applied to supervised image classification.The model consists of a label pyramid and a whole observation field. The potential functions of the coarse grid are derived by simple computations. Then, we propose another scheme introducing a local interaction between two neighbor grids in the label pyramid. This is a way to incorporate cliques with far apart sites for a reasonable price. Finally we present the results on noisy synthetic data and on a SPOT image obtained by different relaxation methods using these models. [bibtex-key = Kato-etal93c]


231. Zoltan Kato, Josiane Zerubia, and Mark Berthod. Satellite Image Classification Using a Modified Metropolis Dynamics. In Proceedings of the International Conference on Acoustics, Speech and Signal Processing, volume 3, San-Francisco, California, USA, pages 573-576, March 1992. IEEE. [PS]
     Abstract: In this paper, we present a pseudo-stochastic variation of the Metropolis dynamics for combinatorial optimization in image classification using Markov Random Fields. At high temperature, the behavior of our algorithm is similar to the stochastic ones. However, if the temperature is less than a certain threshold, it becomes deterministic. The ``length'' of the ``pseudo-stochastic'' phase is controlled by a constant threshold used in the modified dynamics. The algorithm compares favorably to recent stochastic or deterministic methods and yields an approximate but usually good solution to the optimization problem. The algorithm runs on a Connection Machine. It is applied to standard pixel classification problem; objective and subjective comparisons with other algorithms have been made. [bibtex-key = Kato-etal92a]


232. Endre Katona. A Transitive Closure Algorithm for a 16-state Cellprocessor. In Tamas Legendi Gottfried Wolf and Udo Schendel, editors, Proceedings of PARCELLA, volume 342 of Lecture Notes in Computer Science, Berlin, GDR, pages 285-290, 1989. Springer Verlag. [PDF]
     Abstract: Cellprocessors can be considered as microprogrammed Boolean array machines, thus they can process Boolean matrices with very high efficiency. It will be shown that transitive closure of a relation, represented by an nxn Boolean matrix, can be computed in 5n steps using an (n+1)xn array of 16-state cells. If there are several relations, the transitive closure of which should be computed, then a continuous pipeline processing is possible where the processing cost of one matrix is only n steps. The transitive closure algorithm can be partitioned so that arbitrary size relations can be handled with a fixed size cellular array. [bibtex-key = Katona1989]


233. Endre Katona. A Programming Language for Cellular Processors. In Proceedings of Parallel Computing, Berlin, GDR, pages 371-376, 1986. Elsevier (North-Holland). [PDF]
     Abstract: A cellular processor is a homogeneous array of microprogrammable Boolean processors. The cellular array is totally reconfigurable: groups of cells may form arithmetic processing elements (PEs) and these PEs can be interconnected by data channels. This structure can be described by a special graph (called cellular program graph) which is similar to a synchronous circuit model consisting of registers and combinational logic. A language is proposed for the description of such a graph offering a relatively high level programming tool for cellular processors. [bibtex-key = Katona1986]


234. Endre Katona. Assembly-level Programming of Cellular Processors. In Proceedings of PARCELLA, Berlin, GDR, pages 94-100, 1986. Elsevier (North-Holland). [PDF]
     Abstract: We can agree that cell transition functions of a microprogrammed cellular processor form the same level of programming as machine instructions of a conventional computer, because both of them are replaced by microprograms in the course of execution. That is why we say ``assembly level programming'' of a cellular processor if transition functions are described in a given language, the compiler of which can generate an optimal microprogram to each transition function. In this paper cellular assembly languages will be characterized in general, and a cellular assembly program will be given as example for the Cellprocessor Prototype solving multiplication of Boolean matrices. [bibtex-key = Katona1986a]


235. Peter Koles and Endre Katona. Cellular Solution Methods for Different Classes of Picture Processing Algorithms. In Proceedings of Parallel Computing, Berlin, GDR, 1986. Elsevier (North-Holland). [PDF]
     Abstract: The paper gives a classification of picture processing problems considering the input and output data structures and interrelationship between them, and general cellular solution methods are presented for each class. The structural similarity of picture processing algorithms and cellprocessors (both are based on local operations) guarantee the capability of cellprocessors to become powerful devices of picture processing. [bibtex-key = Koeles1986]


236. Endre Katona. String Pattern Matching Algorithms for Cellular Processors. In Proceedings of Parallel Computing, Berlin, GDR, pages 431-436, 1984. Elsevier (North-Holland). [PDF]
     Abstract: Two pattern matching algorithms will be presented, described as software modules for a concrete cellprocessor architecture. The first one is a ``conventional cellular algorithm'' where the patterns are stored in the cellular space, but the second one is based on a special ability of the cellprocessor called ``broadcasting''. Note that our algorithms are suitable for direct hardware implementation, too. Similar VLSI algorithms have been published e.g. by Foster and Kung. However, we prefer the software realization, considering its high flexibility and its capability for interconnection with other cellular software modules. [bibtex-key = Katona1984]


237. Endre Katona. Cellular Algorithms for Binary Matrix Operations. In W. Händler, editor, Proceedings of CONPAR, volume 111 of Lecture Notes in Computer Science, Nurnberg, Germany, pages 203-216, June 1981. Springer Verlag. [PDF]
     Abstract: Binary matrixes (matrixes having only 0 and 1 elements) occur in many theoretical and practical areas (e.g. in the theory of graphs and relations, grammars, coding, etc.) therefore the binary matrix operations have high importance in these topics. It is an obvious idea to use cellular processor consisting of micro-cells for the execution of these operations. The present paper gives a powerful cellular algorithm for the multiplication of two binary matrixes, taking into account the cellprocessor architecture developed by T. Legendi. Moreover, on the basis of this solution cellular algorithms are given for coding and for the computation pf the m-th power of a binary matrix. [bibtex-key = Katona1981]

  This material is presented to ensure timely dissemination of scholarly and technical work. Copyright and all rights therein are retained by authors or by other copyright holders. All person copying this information are expected to adhere to the terms and constraints invoked by each author's copyright. In most cases, these works may not be reposted without the explicit permission of the copyright holder.

 This document was translated from BibTEX by bibtex2html