%0 Book Section %B Proceedings of the ACCV Workshop on Intelligent Mobile and Egocentric Vision (ACCV-IMEV), Lecture Notes in Computer Science %D 2015 %T Collaborative Mobile 3D Reconstruction of Urban Scenes %A Attila Tanacs %A András Majdik %A Levente Hajder %A Jozsef Molnar %A Zsolt Santa %A Zoltan Kato %E Chu-Song Chen %E Mohan Kankanhall %E Shang-Hong Lai %E Joo Hwee %B Proceedings of the ACCV Workshop on Intelligent Mobile and Egocentric Vision (ACCV-IMEV), Lecture Notes in Computer Science %I Springer %C Singapore %P 1-16 %8 Nov 2014 %G eng %9 Conference paper %0 Journal Article %J PATTERN RECOGNITION %D 2015 %T Estimation of linear deformations of 2D and 3D fuzzy objects %A Attila Tanacs %A Joakim Lindbald %A Nataša Sladoje %A Zoltan Kato %X

Registration is a fundamental task in image processing, it is used to determine geometric correspondences between images taken at different times and/or from different viewpoints. Here we propose a general framework in n-dimensions to solve binary shape/object matching problems without the need of establishing additional point or other type of correspondences. The approach is based on generating and solving polynomial systems of equations. We also propose an extension which, provided that a suitable segmentation method can produce a fuzzy border representation, further increases the registration precision. Via numerous synthetic and real test we examine the different solution techniques of the polynomial systems of equations. We take into account a direct analytical, an iterative least-squares, and a combined method. Iterative and combined approaches produce the most precise results. Comparison is made against competing methods for rigid-body problems. Our method is orders of magnitude faster and is able to recover alignment regardless of the magnitude of the deformation compared to the narrow capture range of others. The applicability of the proposed methods is demonstrated on real X-ray images of hip replacement implants and 3D CT volumes of the pelvic area. Since the images must be parsed through only once, our approach is especially suitable for solving registration problems of large images.

%B PATTERN RECOGNITION %I Elsevier %V 48 %P 1387-1399 %8 Apr 2015 %G eng %N 4 %9 Journal Article %R 10.1016/j.patcog.2014.10.006 %0 Journal Article %J Pattern Analysis and Machine Intelligence, IEEE Transactions on %D 2015 %T Realigning 2D and 3D Object Fragments without Correspondences %A Csaba Domokos %A Zoltan Kato %X

This paper addresses the problem of simultaneous estimation of different linear deformations, resulting in a global non-linear transformation, between an original object and its broken fragments. A general framework is proposed without using correspondences, where the solution of a polynomial system of equations directly provides the parameters of the alignment. We quantitatively evaluate the proposed algorithm on a large synthetic dataset containing 2D and 3D images, where linear (rigid-body and affine) transformations are considered. We also conduct an exhaustive analysis of the robustness against segmentation errors and the numerical stability of the proposed method. Moreover, we present experiments on 2D real images as well as on volumetric medical images.

 

%B Pattern Analysis and Machine Intelligence, IEEE Transactions on %I IEEE %V pp %P 1 %8 June 2015 %G eng %N 99 %9 Journal article %R 10.1109/TPAMI.2015.2450726 %0 Conference Paper %B International Conference on Digital Image Computing: Techniques and Applications (DICTA) %D 2014 %T 3D Reconstruction of Planar Patches Seen by Omnidirectional Cameras %A Jozsef Molnar %A Robert Frohlich %A Chetverikov Dmitrij %A Zoltan Kato %E Abdesselam Bouzerdoum %E Lei Wang %E Philip Ogunbona %E Wanqing Li %E Son Lam Phung %B International Conference on Digital Image Computing: Techniques and Applications (DICTA) %I IEEE %C Wollongong, Australia %P 1-8 %G eng %9 Conference paper %0 Book Section %B Proceedings of the ACCV Workshop on Big Data in 3D Computer Vision (ACCV-BigData3DCV) %D 2014 %T 3D Reconstruction of Planar Surface Patches: A Direct Solution %A Jozsef Molnar %A Rui Huang %A Zoltan Kato %E Jian Zhang %E Mohammed Bennamoun %E Fatih Porikli %B Proceedings of the ACCV Workshop on Big Data in 3D Computer Vision (ACCV-BigData3DCV) %I Springer %C Singapore, Szingapúr %P 1-8. %8 Nov 2014 %G eng %9 Conference paper %0 Conference Paper %B International Conference on Pattern Recognition (ICPR) %D 2014 %T Affine Alignment of Occluded Shapes %A Zsolt Santa %A Zoltan Kato %E Michael Felsberg %B International Conference on Pattern Recognition (ICPR) %I IEEE %C Stockholm, Svédország %P 2155-2160 %8 Aug 2014 %@ 978-4-9906441-0-9 %9 Conference paper %0 Conference Paper %B International Conference on Digital Image Computing: Techniques and Applications (DICTA) %D 2014 %T Establishing Correspondences between Planar Image Patches %A Attila Tanacs %A András Majdik %A Jozsef Molnar %A Atul Rai %A Zoltan Kato %E Abdesselam Bouzerdoum %E Lei Wang %E Philip Ogunbona %E Wanqing Li %E Son Lam Phung %B International Conference on Digital Image Computing: Techniques and Applications (DICTA) %I IEEE %C Wollongong, Australia %P 1-7 %8 2014 %G eng %9 Conference paper %0 Book Section %B Informatika a felsőoktatásban 2014 %D 2014 %T Képfeldolgozás a szegedi informatikus-képzésben %A Péter Balázs %A Endre Katona %A Zoltan Kato %A Antal Nagy %A Gábor Németh %A László Gábor Nyúl %A Kálmán Palágyi %A Attila Tanacs %A László Gábor Varga %E Roland Kunkli %E Ildikó Papp %E Edéné Rutkovszky %B Informatika a felsőoktatásban 2014 %I University of Debrecen %C Debrecen, Hungary %P 667-675 %8 2014 %G hun %9 Conference paper %0 Conference Paper %B A Képfeldolgozók és Alakfelismerők Társaságának konferenciája - KÉPAF 2013 %D 2013 %T 2D és 3D bináris objektumok lineáris deformáció-becslésének numerikus megoldási lehetőségei %A Attila Tanacs %A Joakim Lindblad %A Nataša Sladoje %A Zoltan Kato %E László Czúni %B A Képfeldolgozók és Alakfelismerők Társaságának konferenciája - KÉPAF 2013 %I NJSZT-KÉPAF %C Veszprém %P 526 - 541 %8 Jan 2013 %G eng %9 Conference paper %0 Conference Paper %B IEEE Conference on Computer Vision and Pattern Recognition (CVPR) %D 2013 %T Correspondence-less non-rigid registration of triangular surface meshes %A Zsolt Santa %A Zoltan Kato %X

A novel correspondence-less approach is proposed to find a thin plate spline map between a pair of deformable 3D objects represented by triangular surface meshes. The proposed method works without landmark extraction and feature correspondences. The aligning transformation is found simply by solving a system of nonlinear equations. Each equation is generated by integrating a nonlinear function over the object's domains. We derive recursive formulas for the efficient computation of these integrals. Based on a series of comparative tests on a large synthetic dataset, our triangular mesh-based algorithm outperforms state of the art methods both in terms of computing time and accuracy. The applicability of the proposed approach has been demonstrated on the registration of 3D lung CT volumes. © 2013 IEEE.

%B IEEE Conference on Computer Vision and Pattern Recognition (CVPR) %I IEEE %C Portland, OR, USA %P 2275 - 2282 %8 June 2013 %G eng %9 Conference paper %R 10.1109/CVPR.2013.295 %0 Book Section %B Proceedings of International Conference on Digital Image Computing: Techniques and Applications (DICTA) %D 2013 %T Elastic Registration of 3D Deformable Objects %A Zsolt Santa %A Zoltan Kato %E Geoff West %E Péter Kövesi %X

A novel correspondence-less approach is proposed to find a non-linear aligning transformation between a pair of deformable 3D objects. Herein, we consider a polynomial deformation model, but our framework can be easily adapted to other common deformations. The basic idea of the proposed method is to set up a system of nonlinear equations whose solution directly provides the parameters of the aligning transformation. Each equation is generated by integrating a nonlinear function over the object's domains. Thus the number of equations is determined by the number of adopted nonlinear functions yielding a flexible mechanism to generate sufficiently many equations. While classical approaches would establish correspondences between the shapes, our method works without landmarks. The efficiency of the proposed approach has been demonstrated on a large synthetic dataset as well as in the context of medical image registration.

%B Proceedings of International Conference on Digital Image Computing: Techniques and Applications (DICTA) %I IEEE %C New York %P 1 - 7 %8 Nov 2013 %G eng %U http://www.inf.u-szeged.hu/~kato/papers/dicta2012.pdf %9 Conference paper %R 10.1109/DICTA.2012.6411674 %0 Book Section %B Proceedings of the International Symposium on Image and Signal Processing and Analysis (ISPA) %D 2013 %T Evaluation of Point Matching Methods for Wide-baseline Stereo Correspondence on Mobile Platforms %A Endre Juhász %A Attila Tanacs %A Zoltan Kato %E Giovanni Ramponi %E Sven Lončarić %E Alberto Carini %E Karen Egiazarian %X

Wide-baseline stereo matching is a common problem of computer vision. By the explosion of smartphones equipped with camera modules, many classical computer vision solutions have been adapted to such platforms. Considering the widespread use of various networking options for mobile phones, one can consider a set of smart phones as an ad-hoc camera network, where each camera is equipped with a more and more powerful computing engine in addition to a limited bandwidth communication with other devices. Therefore the performance of classical vision algorithms in a collaborative mobile environment is of particular interest. In such a scenario we expect that the images are taken almost simultaneously but from different viewpoints, implying that the camera poses are significantly different but lighting conditions are the same. In this work, we provide quantitative comparison of the most important keypoint detectors and descriptors in the context of wide baseline stereo matching. We found that for resolution of 2 megapixels images the current mobile hardware is capable of providing results efficiently.

 

%B Proceedings of the International Symposium on Image and Signal Processing and Analysis (ISPA) %I IEEE %C Trieste %P 806 - 811 %8 Sep 2013 %G eng %9 Conference paper %R 10.1109/ISPA.2013.6703848 %0 Book Section %B Proceedings of International Joint Conference on Computer Vision, Imaging and Computer Graphics - Theory and Applications (Revised Selected Papers) %D 2013 %T Linear and nonlinear shape alignment without correspondences %A Zoltan Kato %E Paul Richard %E Gabriela Csurka %X

We consider the estimation of diffeomorphic deformations aligning a known binary shape and its distorted observation. The classical solution consists in extracting landmarks, establishing correspondences and then the aligning transformation is obtained via a complex optimization procedure. Herein we present an alternative solution which works without landmark correspondences, is independent of the magnitude of transformation, easy to implement, and has a linear time complexity. The proposed universal framework is capable of recovering linear as well as nonlinear deformations.

 

%B Proceedings of International Joint Conference on Computer Vision, Imaging and Computer Graphics - Theory and Applications (Revised Selected Papers) %S Communications in Computer and Information Science %I Springer Verlag %C Berlin; Heidelberg; New York; London; Paris; Tokyo %P 3 - 17 %8 Feb 2013 %G eng %U http://www.inf.u-szeged.hu/~kato/papers/visapp2012.pdf %9 Conference paper %R 10.1007/978-3-642-38241-3_1 %0 Book Section %B International Conference on Digital Image Computing: Techniques and Applications (DICTA) %D 2013 %T Pose Estimation of Ad-hoc Mobile Camera Networks %A Zsolt Santa %A Zoltan Kato %E Paulo de Souza %E Ulrich Engelke %E Ashfaqur Rahman %X

An algorithm is proposed for the pose estimation of ad-hoc mobile camera networks with overlapping views. The main challenge is to estimate camera parameters with respect to the 3D scene without any specific calibration pattern, hence allowing for a consistent, camera-independent world coordinate system. The only assumption about the scene is that it contains a planar surface patch of a low-rank texture, which is visible in at least two cameras. Such low-rank patterns are quite common in urban environments. The proposed algorithm consists of three main steps: relative pose estimation of the cameras within the network, followed by the localization of the network within the 3D scene using a low-rank surface patch, and finally the estimation of a consistent scale for the whole system. The algorithm follows a distributed architecture, hence the computing power of the participating mobile devices are efficiently used. The performance and robustness of the proposed algorithm have been analyzed on both synthetic and real data. Experimental results confirmed the relevance and applicability of the method.

 

%B International Conference on Digital Image Computing: Techniques and Applications (DICTA) %I IEEE %C Hobart, TAS %P 88 - 95 %8 2013 %G eng %9 Conference paper %M 14000303 %R 10.1109/DICTA.2013.6691514 %0 Book Section %B Proceedings of ICCV Workshop on Big Data in 3D Computer Vision %D 2013 %T Targetless Calibration of a Lidar - Perspective Camera Pair %A Tamás Levente %A Zoltan Kato %E Jian Zhang %E Mohammed Bennamoun %E Dan Schonfeld %E Zhengyou Zhang %X

A novel method is proposed for the calibration of a camera - 3D lidar pair without the use of any special calibration pattern or point correspondences. The proposed method has no specific assumption about the data source: plain depth information is expected from the lidar scan and a simple perspective camera is used for the 2D images. The calibration is solved as a 2D-3D registration problem using a minimum of one (for extrinsic) or two (for intrinsic-extrinsic) planar regions visible in both cameras. The registration is then traced back to the solution of a non-linear system of equations which directly provides the calibration parameters between the bases of the two sensors. The method has been tested on a large set of synthetic lidar-camera image pairs as well as on real data acquired in outdoor environment.

 

%B Proceedings of ICCV Workshop on Big Data in 3D Computer Vision %I IEEE %C Sydney, NSW %P 668 - 675 %8 Dec 2013 %G eng %9 Conference paper %M 14147882 %R 10.1109/ICCVW.2013.92 %0 Book Section %B Intelligent Interactive Technologies and Multimedia %D 2013 %T A unifying framework for correspondence-less shape alignment and its medical applications %A Zoltan Kato %X

We give an overview of our general framework for registering 2D and 3D objects without correspondences. Classical solutions consist in extracting landmarks, establishing correspondences and then the aligning transformation is obtained via a complex optimization procedure. In contrast, our framework works without landmark correspondences, is independent of the magnitude of transformation, easy to implement, and has a linear time complexity. The efficiency and robustness of the method has been demonstarted using various deformations models. Herein, we will focus on medical applications. © 2013 Springer-Verlag.

%B Intelligent Interactive Technologies and Multimedia %S Communications in Computer and Information Science %I Springer %C Allahabad, India %V 276 CCIS %P 40 - 52 %8 March 2013 %@ 1865-0929 %G eng %9 Conference paper %! COMMUN COMPUT INFORM SCI %R 10.1007/978-3-642-37463-0_4 %0 Generic %D 2012 %T Affine Registration of 3D Objects %X

This is the sample implementation and benchmark dataset of the binary image registration algorithm described in the following papers: Attila Tanacs and Zoltan Kato. Fast Linear Registration of 3D Objects Segmented from Medical Images. In Proceedings of International Conference on BioMedical Engineering and Informatics, Shanghai, China, pages 299--303, October 2011. IEEE. Attila Tanacs, Joakim Lindblad, Natasa Sladoje and Zoltan Kato. Estimation of Linear Deformations of 3D Objects. In Proceedings of International Conference on Image Processing, Hong Kong, China, pp. 153-156, September 2010. IEEE.

%8 2012/// %G eng %U http://www.inf.u-szeged.hu/~kato/software/affbin3dregdemo.html %9 Software %0 Book %D 2012 %T Markov random fields in image segmentation %A Zoltan Kato %A Josiane Zerubia %X

Markov Random Fields in Image Segmentation introduces the fundamentals of Markovian modeling in image segmentation as well as providing a brief overview of recent advances in the field.

%I Now Publishers %C Hanover, NH %8 2012 %G eng %9 Book %0 Conference Paper %B International Conference on Pattern Recognition (ICPR) %D 2012 %T A Multi-Layer Phase Field Model for Extracting Multiple Near-Circular Objects %A Csaba Molnar %A Zoltan Kato %A Ian Jermyn %E Jan-Olof Eklundh %E Yuichi Ohta %E Steven Tanimoto %X

This paper proposes a functional that assigns low `energy' to sets of subsets of the image domain consisting of a number of possibly overlapping near-circular regions of approximately a given radius: a `gas of circles'. The model can be used as a prior for object extraction whenever the objects conform to the `gas of circles' geometry, e.g. cells in biological images. Configurations are represented by a multi-layer phase field. Each layer has an associated function, regions being defined by thresholding. Intra-layer interactions assign low energy to configurations consisting of non-overlapping near-circular regions, while overlapping regions are represented in separate layers. Inter-layer interactions penalize overlaps. Here we present a theoretical and experimental analysis of the model.

 

%B International Conference on Pattern Recognition (ICPR) %I IEEE %C Tsukuba, Japan %P 1427 - 1430 %8 Nov 2012 %@ 978-1-4673-2216-4 %G eng %9 Conference paper %M 13324819 %0 Journal Article %J IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE %D 2012 %T Nonlinear Shape Registration without Correspondences %A Csaba Domokos %A Jozsef Nemeth %A Zoltan Kato %X

In this paper, we propose a novel framework to estimate the parameters of a diffeomorphism that aligns a known shape and its distorted observation. Classical registration methods first establish correspondences between the shapes and then compute the transformation parameters from these landmarks. Herein, we trace back the problem to the solution of a system of nonlinear equations which directly gives the parameters of the aligning transformation. The proposed method provides a generic framework to recover any diffeomorphic deformation without established correspondences. It is easy to implement, not sensitive to the strength of the deformation, and robust against segmentation errors. The method has been applied to several commonly used transformation models. The performance of the proposed framework has been demonstrated on large synthetic data sets as well as in the context of various applications.

 

%B IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE %I IEEE %V 34 %P 943 - 958 %8 2012 %@ 0162-8828 %G eng %U http://www.inf.u-szeged.hu/~kato/papers/TPAMI-2010-03-0146.R2_Kato.pdf %N 5 %9 Journal article %M 12617610 %! IEEE T PATTERN ANAL %R 10.1109/TPAMI.2011.200 %0 Book Section %B Advanced color image processing and analysis %D 2012 %T Parametric Stochastic Modeling for Color Image Segmentation and Texture Characterization %A Imtnan-Ul-Haque Qazi %A Oliver Alata %A Zoltan Kato %E Christine Fernandez-Maloigne %X

Black should be made a color of light Clemence Boulouque

Parametric stochastic models offer the definition of color and/or texture features based on model parameters, which is of interest for color texture classification, segmentation and synthesis.

In this chapter, distribution of colors in the images through various parametric approximations including multivariate Gaussian distribution, multivariate Gaussian mixture models (MGMM) and Wishart distribution, is discussed. In the context of Bayesian color image segmentation, various aspects of sampling from the posterior distributions to estimate the color distribution from MGMM and the label field, using different move types are also discussed. These include reversible jump mechanism from MCMC methodology. Experimental results on color images are presented and discussed.

Then, we give some materials for the description of color spatial structure using Markov Random Fields (MRF), and more particularly multichannel GMRF, and multichannel linear prediction models. In this last approach, two dimensional complex multichannel versions of both causal and non-causal models are discussed to perform the simultaneous parametric power spectrum estimation of the luminance and the chrominance channels of the color image. Application of these models to the classification and segmentation of color texture images is also illustrated.

 

%B Advanced color image processing and analysis %I Springer %C Berlin; Heidelberg; New York; London; Paris; Tokyo %P 279 - 325 %8 2012 %@ 978-1-4419-6189-1 %G eng %9 Book chapter %R 10.1007/978-1-4419-6190-7_9 %0 Conference Paper %B International Conference on Pattern Recognition (ICPR) %D 2012 %T Simultaneous Affine Registration of Multiple Shapes %A Csaba Domokos %A Zoltan Kato %E Jan-Olof Eklundh %E Yuichi Ohta %E Steven Tanimoto %X

The problem of simultaneously estimating affine deformations between multiple objects occur in many applications. Herein, a direct method is proposed which provides the result as a solution of a linear system of equations without establishing correspondences between the objects. The key idea is to construct enough linearly independent equations using covariant functions, and then finding the solution simultaneously for all affine transformations. Quantitative evaluation confirms the performance of the method.

 

%B International Conference on Pattern Recognition (ICPR) %I IEEE %C Tsukuba, Japan %P 9 - 12 %8 Nov 2012 %@ 978-1-4673-2216-4 %G eng %9 Conference paper %M 13324478 %0 Conference Paper %B International Conference on Pattern Recognition (ICPR) %D 2012 %T Spectral clustering to model deformations for fast multimodal prostate registration %A Jhimli Mitra %A Zoltan Kato %A Soumya Ghose %A Desire Sidibe %A Robert Martí %A Xavier Lladó %A Oliver Arnau %A Joan C Vilanova %A Fabrice Meriaudeau %E Jan-Olof Eklundh %E Yuichi Ohta %E Steven Tanimoto %X

This paper proposes a method to learn deformation parameters off-line for fast multimodal registration of ultrasound and magnetic resonance prostate images during ultrasound guided needle biopsy. The registration method involves spectral clustering of the deformation parameters obtained from a spline-based nonlinear diffeomorphism between training magnetic resonance and ultrasound prostate images. The deformation models built from the principal eigen-modes of the clusters are then applied on a test magnetic resonance image to register with the test ultrasound prostate image. The deformation model with the least registration error is finally chosen as the optimal model for deformable registration. The rationale behind modeling deformations is to achieve fast multimodal registration of prostate images while maintaining registration accuracies which is otherwise computationally expensive. The method is validated for 25 patients each with a pair of corresponding magnetic resonance and ultrasound images in a leave-one-out validation framework. The average registration accuracies i.e. Dice similarity coefficient of 0.927 ± 0.025, 95% Hausdorff distance of 5.14 ± 3.67 mm and target registration error of 2.44 ± 1.17 mm are obtained by our method with a speed-up in computation time by 98% when compared to Mitra et al. [7].

 

%B International Conference on Pattern Recognition (ICPR) %I IEEE %C Tsukuba, Japan %P 2622 - 2625 %8 Nov 2012 %@ 978-1-4673-2216-4 %G eng %U http://hal.archives-ouvertes.fr/docs/00/71/09/43/PDF/ICPR_Jhimli.pdf %9 Conference paper %M 13325059 %0 Journal Article %J MEDICAL IMAGE ANALYSIS %D 2012 %T A spline-based non-linear diffeomorphism for multimodal prostate registration. %A Jhimli Mitra %A Zoltan Kato %A Robert Martí %A Oliver Arnau %A Xavier Lladó %A Desire Sidibe %A Soumya Ghose %A Joan C Vilanova %A Josep Comet %A Fabrice Meriaudeau %X

This paper presents a novel method for non-rigid registration of transrectal ultrasound and magnetic resonance prostate images based on a non-linear regularized framework of point correspondences obtained from a statistical measure of shape-contexts. The segmented prostate shapes are represented by shape-contexts and the Bhattacharyya distance between the shape representations is used to find the point correspondences between the 2D fixed and moving images. The registration method involves parametric estimation of the non-linear diffeomorphism between the multimodal images and has its basis in solving a set of non-linear equations of thin-plate splines. The solution is obtained as the least-squares solution of an over-determined system of non-linear equations constructed by integrating a set of non-linear functions over the fixed and moving images. However, this may not result in clinically acceptable transformations of the anatomical targets. Therefore, the regularized bending energy of the thin-plate splines along with the localization error of established correspondences should be included in the system of equations. The registration accuracies of the proposed method are evaluated in 20 pairs of prostate mid-gland ultrasound and magnetic resonance images. The results obtained in terms of Dice similarity coefficient show an average of 0.980+/-0.004, average 95% Hausdorff distance of 1.63+/-0.48mm and mean target registration and target localization errors of 1.60+/-1.17mm and 0.15+/-0.12mm respectively.

%B MEDICAL IMAGE ANALYSIS %V 16 %P 1259 - 1279 %8 Aug 2012 %@ 1361-8415 %G eng %N 6 %9 Journal article %! MED IMAGE ANAL %R 10.1016/j.media.2012.04.006 %0 Book Section %B International Conference on Image Analysis and Recognition (ICIAR) %D 2012 %T A Unifying Framework for Correspondence-less Linear Shape Alignment %A Zoltan Kato %E Aurélio Campilho %X

We consider the estimation of linear 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. Here we propose a unified framework where the exact transformation is obtained as the solution of either a polynomial or a linear system of equations without establishing correspondences. The advantages of the proposed solutions are that they are fast, easy to implement, have linear time complexity, work without landmark correspondences and are independent of the magnitude of transformation.

 

%B International Conference on Image Analysis and Recognition (ICIAR) %S Lecture Notes in Computer Science %I Springer Verlag %C Aveiro, Portugal %P 277 - 284 %8 June 2012 %@ 978-3-642-31294-6 %G eng %9 Conference paper %! LNCS %R 10.1007/978-3-642-31295-3_33 %0 Conference Paper %B IEEE International Conference on Cognitive Infocommunications (CogInfoCom) %D 2012 %T A Unifying Framework for Non-linear Registration of 3D Objects %A Zsolt Santa %A Zoltan Kato %X

An extension of our earlier work is proposed to find a non-linear aligning transformation between a pair of deformable 3D objects. The basic idea is to set up a system of nonlinear equations whose solution directly provides the parameters of the aligning transformation. Each equation is generated by integrating a nonlinear function over the object's domains. Thus the number of equations is determined by the number of adopted nonlinear functions yielding a flexible mechanism to generate sufficiently many equations. While classical approaches would establish correspondences between the shapes, our method works without landmarks. Experiments with 3D polynomial and thin plate spline deformations confirm the performance of the framework.

 

%B IEEE International Conference on Cognitive Infocommunications (CogInfoCom) %I IEEE %C Kosice, Slovakia %P 547 - 552 %8 Dec 2012 %@ 978-1-4673-5187-4 %G eng %U http://www.inf.u-szeged.hu/~kato/papers/coginfocomm2012.pdf %9 Conference paper %R 10.1109/CogInfoCom.2012.6422041 %0 Conference Paper %B A Képfeldolgozók és Alakfelismerők Társaságának konferenciája - KÉPAF 2011 %D 2011 %T 3D objektumok lineáris deformációinak becslése %A Attila Tanacs %A Joakim Lindblad %A Nataša Sladoje %A Zoltan Kato %E Zoltan Kato %E Kálmán Palágyi %B A Képfeldolgozók és Alakfelismerők Társaságának konferenciája - KÉPAF 2011 %I NJSZT %C Szeged %P 471 - 480 %8 Jan 2011 %G eng %9 Conference paper %0 Conference Paper %B A Képfeldolgozók és Alakfelismerők Társaságának konferenciája - KÉPAF 2011 %D 2011 %T Affin Puzzle: Deformált objektumdarabok helyreállítása megfeleltetések nélkül %A Csaba Domokos %A Zoltan Kato %E Zoltan Kato %E Kálmán Palágyi %B A Képfeldolgozók és Alakfelismerők Társaságának konferenciája - KÉPAF 2011 %I NJSZT %C Szeged %P 206 - 220 %8 Jan 2011 %G eng %U http://www.inf.u-szeged.hu/kepaf2011/pdfs/S05_03.pdf %9 Conference paper %0 Conference Paper %B A Képfeldolgozók és Alakfelismerők Társaságának konferenciája - KÉPAF 2011 %D 2011 %T Bináris tomográfiai rekonstrukció objektum alapú evolúciós algoritmussal %B A Képfeldolgozók és Alakfelismerők Társaságának konferenciája - KÉPAF 2011 %I NJSZT %C Szeged %P 117 - 127 %8 Jan 2011 %G eng %9 Conference paper %0 Conference Paper %B A Képfeldolgozók és Alakfelismerők Társaságának konferenciája - KÉPAF 2011 %D 2011 %T Élősejt szegmentálása gráfvágás segítségével fluoreszcenciás mikroszkóp képeken %A Milan Lesko %A Zoltan Kato %A Antal Nagy %A Imre Gombos %A Zsolt Török %A László Vígh %A László Vígh %E Zoltan Kato %E Kálmán Palágyi %B A Képfeldolgozók és Alakfelismerők Társaságának konferenciája - KÉPAF 2011 %I NJSZT %C Szeged %P 319 - 328 %8 Jan 2011 %G eng %U http://www.inf.u-szeged.hu/kepaf2011/pdfs/S08_02.pdf %9 Conference paper %0 Conference Paper %B Biomedical Engineering and Informatics (BMEI) %D 2011 %T Fast linear registration of 3D objects segmented from medical images %A Attila Tanacs %A Zoltan Kato %E Yongsheng Ding %E Yonghong Peng %E Riyi Shi %E Kuangrong Hao %E Lipo Wang %X

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. © 2011 IEEE.

%B Biomedical Engineering and Informatics (BMEI) %I IEEE %C Shanghai %P 294 - 298 %8 Oct 2011 %@ 978-1-4244-9351-7 %G eng %9 Conference paper %M 12436502 %R 10.1109/BMEI.2011.6098290 %0 Conference Paper %B A Képfeldolgozók és Alakfelismerők Társaságának konferenciája - KÉPAF 2011 %D 2011 %T Iterációnkénti simítással kombinált vékonyítás %A Péter Kardos %A Gábor Németh %A Kálmán Palágyi %E Zoltan Kato %E Kálmán Palágyi %B A Képfeldolgozók és Alakfelismerők Társaságának konferenciája - KÉPAF 2011 %I NJSZT %C Szeged %P 174 - 189 %8 Jan 2011 %G eng %U http://www.inf.u-szeged.hu/kepaf2011/pdfs/S05_01.pdf %9 Conference paper %0 Conference Paper %B A Képfeldolgozók és Alakfelismerők Társaságának konferenciája - KÉPAF 2011 %D 2011 %T Mediánszűrés alkalmazása algebrai rekonstrukciós módszerekben %A Norbert Hantos %A Péter Balázs %E Zoltan Kato %E Kálmán Palágyi %B A Képfeldolgozók és Alakfelismerők Társaságának konferenciája - KÉPAF 2011 %I NJSZT %C Szeged %P 106 - 116 %8 Jan 2011 %G eng %9 Conference paper %0 Book Section %B Advances Concepts for Intelligent Vision Systems (ACIVS) %D 2011 %T A Multi-Layer 'Gas of Circles' Markov Random Field Model for the Extraction of Overlapping Near-Circular Objects %X

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.

%B Advances Concepts for Intelligent Vision Systems (ACIVS) %S Lecture Notes in Computer Science %I Springer-Verlag %C Ghent, Belgium %P 171 - 182 %8 Aug 2011 %@ 978-3-642-23686-0 %G eng %U http://www.inf.u-szeged.hu/ipcg/publications/Year/2011.complete.xml#Nemeth-etal2011 %9 Conference paper %! LNCS %R 10.1007/978-3-642-23687-7_16 %0 Conference Paper %B International Conference on Digital Image Computing: Techniques and Applications (DICTA) %D 2011 %T A non-linear diffeomorphic framework for prostate multimodal registration %A Jhimli Mitra %A Zoltan Kato %A Robert Martí %A Oliver Arnau %A Xavier Lladó %A Soumya Ghose %A Joan C Vilanova %A Fabrice Meriaudeau %X

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. © 2011 IEEE.

%B International Conference on Digital Image Computing: Techniques and Applications (DICTA) %I IEEE %C Noosa, QLD %P 31 - 36 %8 Dec 2011 %@ 978-1-4577-2006-2 %G eng %9 Conference paper %M 12476651 %R 10.1109/DICTA.2011.14 %0 Generic %D 2011 %T Nonlinear Shape Registration without Correspondences %A Zoltán Kornél Török %A Csaba Domokos %A Jozsef Nemeth %A Zoltan Kato %X

This is the sample implementation and benchmark dataset of the nonlinear registration of 2D shapes described in the following papers: Csaba Domokos, Jozsef Nemeth, and Zoltan Kato. Nonlinear Shape Registration without Correspondences. IEEE Transactions on Pattern Analysis and Machine Intelligence, 34(5):943--958, May 2012. Note that the current demo program implements only planar homography deformations. Other deformations can be easily implemented based on the demo code.

%8 2011/// %G eng %U http://www.inf.u-szeged.hu/~kato/software/planarhombinregdemo.html %9 Software %0 Book %D 2011 %T Számítógépes látás %A Zoltan Kato %A László Czúni %I Typotex Kiadó %C Budapest %8 2011 %G hun %9 Book %0 Conference Paper %B A Képfeldolgozók és Alakfelismerők Társaságának konferenciája - KÉPAF 2011 %D 2011 %T A topológia-megőrzés elegendő feltételein alapuló 3D párhuzamos vékonyító algoritmusok %A Gábor Németh %A Péter Kardos %A Kálmán Palágyi %E Zoltan Kato %E Kálmán Palágyi %B A Képfeldolgozók és Alakfelismerők Társaságának konferenciája - KÉPAF 2011 %I NJSZT %C Szeged %P 190 - 205 %8 Jan 2011 %G hun %U http://www.inf.u-szeged.hu/kepaf2011/pdfs/S05_02.pdf %9 Conference paper %0 Conference Paper %B A Képfeldolgozók és Alakfelismerők Társaságának konferenciája - KÉPAF 2011 %D 2011 %T Vetületi irányfüggőség a bináris tomográfiában %A László Gábor Varga %A Péter Balázs %A Antal Nagy %E Zoltan Kato %E Kálmán Palágyi %B A Képfeldolgozók és Alakfelismerők Társaságának konferenciája - KÉPAF 2011 %I NJSZT %C Szeged %P 92 - 105 %8 Jan 2011 %G hun %9 Conference paper %0 Book Section %B European Conference on Computer Vision (ECCV) %D 2010 %T Affine puzzle: Realigning deformed object fragments without correspondences %A Csaba Domokos %A Zoltan Kato %E Kostas Daniilidis %E Petros Maragos %E Nikos Paragios %X

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. © 2010 Springer-Verlag.

%B European Conference on Computer Vision (ECCV) %S Lecture Notes in Computer Science %I Springer %C Crete, Greece %P 777 - 790 %8 Sep 2010 %@ 978-3-642-15551-2 %G eng %9 Conference paper %! LNCS %R 10.1007/978-3-642-15552-9_56 %0 Book Section %B IEEE International Conference on Image Processing (ICIP) %D 2010 %T Estimation of linear deformations of 3D objects %A Attila Tanacs %A Joakim Lindblad %A Nataša Sladoje %A Zoltan Kato %X

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 e.g. 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. © 2010 IEEE.

%B IEEE International Conference on Image Processing (ICIP) %I IEEE %C Hong Kong, Hong Kong %P 153 - 156 %8 Sep 2010 %G eng %9 Conference paper %0 Book Section %B 20th international conference on pattern recognition (ICPR 2010) %D 2010 %T Live cell segmentation in fluorescence microscopy via graph cut %A Milan Lesko %A Zoltan Kato %A Antal Nagy %A Imre Gombos %A Zsolt Török %A László Vígh %A László Vígh %E Aytul Ercil %X

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. © 2010 IEEE.

%B 20th international conference on pattern recognition (ICPR 2010) %I IEEE %C Istanbul, Turkey %P 1485 - 1488 %8 Aug 2010 %@ 978-1-4244-7542-1 %G eng %9 Conference paper %M 11593484 %R 10.1109/ICPR.2010.367 %0 Journal Article %J PATTERN RECOGNITION %D 2010 %T Parametric estimation of affine deformations of planar shapes %A Csaba Domokos %A Zoltan Kato %B PATTERN RECOGNITION %V 43 %P 569 - 578 %8 March 2010 %@ 0031-3203 %G eng %N 3 %9 Journal article %! PATTERN RECOGN %R 10.1016/j.patcog.2009.08.013 %0 Book Section %B Proceedings of the 6th International Conference on Signal Image Technology and Internet Based Systems, SITIS 2010 %D 2010 %T SITIS 2010: Track SIT editorial message: Signal and Image Technologies %A Albert Dipanda %A Zoltan Kato %E Albert Dipanda %E Richard Chbeir %E Kokou Yetongnon %B Proceedings of the 6th International Conference on Signal Image Technology and Internet Based Systems, SITIS 2010 %I IEEE Computer Society Press %C Kuala Lumpur %P XV %8 2010 %G eng %0 Book Section %B 16th IEEE International Conference on Image Processing (ICIP), 2009 %D 2009 %T Affine alignment of compound objects: A direct approach %A Csaba Domokos %A Zoltan Kato %X

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-of-the-art methods both in terms of precision, robustness and complexity. ©2009 IEEE.

%B 16th IEEE International Conference on Image Processing (ICIP), 2009 %I IEEE %C Cairo, Egypt %P 169 - 172 %8 Nov 2009 %@ 978-1-4244-5653-6 %G eng %9 Conference paper %M 11150920 %R 10.1109/ICIP.2009.5414195 %0 Generic %D 2009 %T Affine Registration of Planar Shapes %A Zsolt Katona %A Csaba Domokos %A Zoltan Kato %X

This is the sample implementation and benchmark dataset of the binary image registration algorithm described in the following paper: Csaba Domokos and Zoltan Kato. Parametric Estimation of Affine Deformations of Planar Shapes. Pattern Recognition, 43(3):569--578, March 2010.

%8 2009/// %G eng %U http://www.inf.u-szeged.hu/~kato/software/affbinregdemo.html %0 Journal Article %J IEEE TRANSACTIONS ON IMAGE PROCESSING %D 2009 %T Detection of Object Motion Regions in Aerial Image Pairs with a Multilayer Markovian Model %A Csaba Benedek %A Tamas Sziranyi %A Zoltan Kato %A Josiane Zerubia %X

We propose a new Bayesian method for detectingthe regions of object displacements in aerial image pairs. We use a robust but coarse 2-D 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 (L3MRF) model which integrates information from two different features, and ensures connected homogenous regions in the segmented images. Validation is given on real aerial photos.

%B IEEE TRANSACTIONS ON IMAGE PROCESSING %I IEEE %V 18 %P 2303 - 2315 %8 2009 %@ 1057-7149 %G eng %N 10 %9 Journal article %! IEEE T IMAGE PROCESS %R 10.1109/TIP.2009.2025808 %0 Journal Article %J PATTERN RECOGNITION %D 2009 %T A higher-order active contour model of a 'gas of circles' and its application to tree crown extraction %A Peter Horvath %A Ian Jermyn %A Zoltan Kato %A Josiane Zerubia %B PATTERN RECOGNITION %V 42 %P 699 - 709 %8 2009/// %@ 0031-3203 %G eng %N 5 %! PATTERN RECOGN %0 Conference Paper %B A Képfeldolgozók és Alakfelismerők Társaságának konferenciája - KÉPAF 2009 %D 2009 %T Kör alakú objektumok szegmentálása Markov mező segítségével %B A Képfeldolgozók és Alakfelismerők Társaságának konferenciája - KÉPAF 2009 %I Akaprint %C Budapest %P 1 - 9 %8 Jan 2009 %G hun %U http://vision.sztaki.hu/~kepaf/kepaf2009_CD/files/116-4-MRFCircle08.pdf %9 Conference paper %0 Book Section %B 16th IEEE International Conference on Image Processing (ICIP) %D 2009 %T A Markov random field model for extracting near-circular shapes %A Tamás Blaskovics %A Zoltan Kato %A Ian Jermyn %X

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 resultingMRF 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. ©2009 IEEE.

%B 16th IEEE International Conference on Image Processing (ICIP) %I IEEE %C Cairo, Egypt %P 1073 - 1076 %8 Nov 2009 %@ 978-1-4244-5653-6 %G eng %9 Conference paper %R 10.1109/ICIP.2009.5413472 %0 Conference Paper %B 16th IEEE International Conference on Image Processing (ICIP) %D 2009 %T Nonlinear registration of binary shapes %A Jozsef Nemeth %A Csaba Domokos %A Zoltan Kato %X

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 non-linear 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. ©2009 IEEE.

%B 16th IEEE International Conference on Image Processing (ICIP) %I IEEE %C Cairo, Egypt %P 1101 - 1104 %8 Nov 2009 %@ 978-1-4244-5653-6 %G eng %9 Conference paper %R 10.1109/ICIP.2009.5413468 %0 Book Section %B Image Analysis %D 2009 %T Recovering affine deformations of fuzzy shapes %A Attila Tanacs %A Csaba Domokos %A Nataša Sladoje %A Joakim Lindblad %A Zoltan Kato %E Arnt-Borre Salberg %E Jon Yngve Hardeberg %E Robert Jenssen %X

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. © 2009 Springer Berlin Heidelberg.

%B Image Analysis %S Lecture Notes in Computer Science %I Springer-Verlag %C Oslo, Norway %P 735 - 744 %8 June 2009 %G eng %9 Conference paper %! LNCS %R 10.1007/978-3-642-02230-2_75 %0 Book Section %B 12th International Conference on Computer Vision, ICCV 2009 %D 2009 %T Recovering planar homographies between 2D shapes %X

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. ©2009 IEEE.

%B 12th International Conference on Computer Vision, ICCV 2009 %I IEEE %P 2170 - 2176 %8 2009/// %G eng %0 Conference Paper %B A Képfeldolgozók és Alakfelismerők Társaságának konferenciája - KÉPAF 2009 %D 2009 %T Síkbeli alakzatok regisztrációja kovariáns függvények felhasználásával %B A Képfeldolgozók és Alakfelismerők Társaságának konferenciája - KÉPAF 2009 %I Akaprint %C Budapest %P 1 - 8 %8 Jan 2009 %G eng %9 Conference papers %0 Conference Paper %B A Képfeldolgozók és Alakfelismerők Társaságának konferenciája - KÉPAF 2009 %D 2009 %T Síkhomográfia paramétereinek becslése bináris képeken %B A Képfeldolgozók és Alakfelismerők Társaságának konferenciája - KÉPAF 2009 %I Akaprint %C Budapest %P 1 - 8 %8 Jan 2009 %G eng %9 Conference paper %0 Generic %D 2009 %T Supervised Color Image Segmentation in a Markovian Framework %A Mihály Gara %A Zoltan Kato %X

This is the sample implementation of a Markov random field based color image segmentation algorithm described in the following paper: 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. Note that the current demo program implements only a supervised version of the segmentation method described in the above paper (i.e. parameter values are learned interactively from representative regions selected by the user). Otherwise, the program implements exactly the color MRF model proposed in the paper. Images are automatically converted from RGB to the perceptually uniform CIE-L*u*v* color space before segmentation.

%8 2009/// %G eng %U http://www.inf.u-szeged.hu/~kato/software/colormrfdemo.html %0 Book Section %B Image Analysis and Recognition %D 2008 %T Binary image registration using covariant gaussian densities %A Csaba Domokos %A Zoltan Kato %E Aurélio Campilho %X

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. © 2008 Springer-Verlag Berlin Heidelberg.

%B Image Analysis and Recognition %S Lecture Notes in Computer Science %I Springer %C Póvoa de Varzim, Portugal %P 455 - 464 %8 June 2008 %@ 978-3-540-69811-1 %G eng %9 Conference paper %! LNCS %R 10.1007/978-3-540-69812-8_45 %0 Book Section %B Informatika a felsőoktatásban 2008 %D 2008 %T A képfeldolgozás kutatása a Szegedi Tudományegyetemen %X A digitális képfeldolgozás kutatásának a Szegedi TudományegyetemTermészettudományi és Informatikai Karán, az Informatikai Tanszékcsoport Képfeldolgozás és Számítógépes Grafika Tanszékén közel négy évtizedes hagyománya van. A Tanszék valamennyi munkatársa nemzetközileg elismert kutatómunkát folytat, melyet már több száz rangos publikáció fémjelez. Számos, a képfeldolgozás kutatásában vezető egyetemmel és kutatóintézettel építettünk ki szoros kapcsolatot és folytattunk eredményes kutatómunkát, aktív résztvevői vagyunk a hazai és a nemzetközi tudományos közéletnek. A legfontosabb, jelenleg is folyó kutatásaink: orvosi képek feldolgozása, diszkrét tomográfia, képszegmentálás, térinformatika, távérzékelés, képregisztráció, vázkijelölés, műtéti tervezés. %B Informatika a felsőoktatásban 2008 %I Debreceni Egyetem Informatikai Kar %C Debrecen %8 2008/// %G eng %U http://www.agr.unideb.hu/if2008/kiadvany/papers/E62.pdf %0 Book Section %B Proceedings of the IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) %D 2008 %T Parametric estimation of affine deformations of binary images %A Csaba Domokos %A Zoltan Kato %A Joseph M Francos %X

We consider the problem of planar object registration on binary images where the aligning transformation is restricted to the group of 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 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. ©2008 IEEE.

%B Proceedings of the IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) %I IEEE %C Las Vegas, NV, USA %P 889 - 892 %8 March 2008 %@ 978-1-4244-1483-3 %G eng %9 Conference paper %M 9973096 %R 10.1109/ICASSP.2008.4517753 %0 Journal Article %J IMAGE AND VISION COMPUTING %D 2008 %T Segmentation of color images via reversible jump MCMC sampling %A Zoltan Kato %B IMAGE AND VISION COMPUTING %I Elsevier %V 26 %P 361 - 371 %8 March 2008 %@ 0262-8856 %G eng %N 3 %9 Journal article %! IMAGE VISION COMPUT %R 10.1016/j.imavis.2006.12.004 %0 Conference Paper %B A Képfeldolgozók és Alakfelismerők Társaságának konferenciája - KÉPAF 2007 %D 2007 %T Kör alakú objektumok szegmentálása magasabb rendű aktív kontúr modellek segítségével %B A Képfeldolgozók és Alakfelismerők Társaságának konferenciája - KÉPAF 2007 %I Képfeldolgozók és Alakfelismerők Társasága %C Debrecen %P 133 - 140 %8 Jan 2007 %G eng %9 Conference paper %0 Thesis %D 2007 %T Markovian Image Models and their Application in Image Segmentation %A Zoltan Kato %8 2007 %G eng %9 PhD Thesis %0 Conference Paper %B A Képfeldolgozók és Alakfelismerők Társaságának konferenciája - KÉPAF 2007 %D 2007 %T Parametric Estimation of Two-Dimensional Affine Transformations of Binary Images %B A Képfeldolgozók és Alakfelismerők Társaságának konferenciája - KÉPAF 2007 %I Képfeldolgozók és Alakfelismerők Társasága %C Debrecen %P 257 - 265 %8 Jan 2007 %G eng %9 Conference paper %0 Generic %D 2007 %T A Three-layer MRF model for Object Motion Detection in Airborne Images %A Csaba Benedek %A Tamas Sziranyi %A Zoltan Kato %A Josiane Zerubia %8 2007/// %G eng %0 Book Section %B Proceedings of the18th International Conference on Pattern Recognition, ICPR 2006 %D 2006 %T A higher-order active contour model for tree detection %X

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. © 2006 IEEE.

%B Proceedings of the18th International Conference on Pattern Recognition, ICPR 2006 %I IEEE %P 130 - 133 %8 2006/// %G eng %0 Conference Paper %B Proceedings of the International Conference on Pattern Recognition (ICPR) %D 2006 %T A Higher-Order Active Contour Model for Tree Detection %A Peter Horvath %A Ian Jermyn %A Zoltan Kato %A Josiane Zerubia %X

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.

%B Proceedings of the International Conference on Pattern Recognition (ICPR) %I IAPR %C Hong Kong, China %V 2 %P 130–133 %8 2006 %9 Conference paper %0 Generic %D 2006 %T A Higher-Order Active Contour Model of a `Gas of Circles' and its Application to Tree Crown Extraction %8 2006/// %G eng %0 Book Section %B Proceedings of Indian Conference on Computer Vision, Graphics and Image Processing (ICVGIP) %D 2006 %T An Improved `Gas of Circles' Higher-Order Active Contour Model and its Application to Tree Crown Extraction %B Proceedings of Indian Conference on Computer Vision, Graphics and Image Processing (ICVGIP) %I Springer Verlag %C Berlin; Heidelberg; New York %P 152 - 161 %8 2006/// %G eng %0 Journal Article %J IMAGE AND VISION COMPUTING %D 2006 %T A Markov random field image segmentation model for color textured images %B IMAGE AND VISION COMPUTING %V 24 %P 1103 - 1114 %8 2006/// %@ 0262-8856 %G eng %N 10 %! IMAGE VISION COMPUT %0 Book Section %B Proceedings - 14th International Conference on Image Processing, ICIP 2007 %D 2006 %T A multi-layer MRF model for object-motion detection in unregistered airborne image-pairs %B Proceedings - 14th International Conference on Image Processing, ICIP 2007 %I IEEE %C Piscataway %P VI-141 - VI-144 %8 2006/// %G eng %U http://www.icip2007.org/Papers/AcceptedList.asp %0 Book Section %B COMPUTER VISION - ACCV 2006, PT II %D 2006 %T A multi-layer MRF model for video object segmentation %B COMPUTER VISION - ACCV 2006, PT II %I Springer Verlag %P 953 - 962 %8 2006/// %G eng %0 Book Section %B Joint Hungarian-Austrian conference on image processing and pattern recognition. 5th conference of the Hungarian Association for Image Processing and Pattern Recognition (KÉPAF), 29th workshop of the Austrian Association for Pattern Reco %D 2005 %T Shape Moments for Region Based Active Contours %B Joint Hungarian-Austrian conference on image processing and pattern recognition. 5th conference of the Hungarian Association for Image Processing and Pattern Recognition (KÉPAF), 29th workshop of the Austrian Association for Pattern Reco %I OCG %C Vienna %P 187 - 194 %8 2005/// %G eng %0 Generic %D 2005 %T Supervised Image Segmentation Using Markov Random Fields %X This is the sample implementation of a Markov random field based image segmentation algorithm described in the following papers: Mark Berthod, Zoltan Kato, Shan Yu, and Josiane Zerubia. Bayesian Image Classification Using Markov Random Fields. Image and Vision Computing, 14:285--295, 1996. Keyword(s): Bayesian image classification, Markov random fields, Optimisation. Zoltan Kato, Josiane Zerubia, and Mark Berthod. Satellite Image Classification Using a Modified Metropolis Dynamics. In Proceedings of International Conference on Acoustics, Speech and Signal Processing, volume 3, San-Francisco, California, USA, pages 573-576, March 1992. IEEE. Zoltan Kato. Modélisations markoviennes multirésolutions en vision par ordinateur. Application a` la segmentation d'images SPOT. PhD Thesis, INRIA, Sophia Antipolis, France, December 1994. Note: Available in French (follow the URL link) and English. Keyword(s): computer vision, early vision, Markovian model, multiscale model, hierarchical model, parallel combinatorial optimization algorithm, multi-temperature annealing, parameter estimation. %8 2005/// %G eng %U http://www.inf.u-szeged.hu/~kato/software/mrfdemo.html %0 Book Section %B Joint Hungarian-Austrian conference on image processing and pattern recognition. 5th conference of the Hungarian Association for Image Processing and Pattern Recognition (KÉPAF), 29th workshop of the Austrian Association for Pattern Reco %D 2005 %T Video Object Segmentation Using a Multicue Markovian Model %B Joint Hungarian-Austrian conference on image processing and pattern recognition. 5th conference of the Hungarian Association for Image Processing and Pattern Recognition (KÉPAF), 29th workshop of the Austrian Association for Pattern Reco %I OCG %C Vienna %P 111 - 118 %8 2005/// %G eng %0 Conference Paper %B A Képfeldolgozók és Alakfelismerők Társaságának konferenciája - KÉPAF 2004 %D 2004 %T Color, Texture and Motion Segmentation Using Gradient Vector Flow %B A Képfeldolgozók és Alakfelismerők Társaságának konferenciája - KÉPAF 2004 %C Miskolctapolca %P 131 - 137 %8 Jan 2004 %G eng %0 Conference Paper %B A Képfeldolgozók és Alakfelismerők Társaságának konferenciája - KÉPAF 2004 %D 2004 %T Color textured image segmentation using a multi-layer Markovian model %B A Képfeldolgozók és Alakfelismerők Társaságának konferenciája - KÉPAF 2004 %C Miskolctapolca %P 152 - 158 %8 Jan 2004 %G eng %0 Conference Paper %B A Képfeldolgozók és Alakfelismerők Társaságának konferenciája - KÉPAF 2004 %D 2004 %T Optical Flow Computation Using an Energy Minimization Approach %B A Képfeldolgozók és Alakfelismerők Társaságának konferenciája - KÉPAF 2004 %C Miskolctapolca %P 125 - 130 %8 Jan 2004 %G eng %0 Conference Paper %D 2004 %T Reversible Jump Markov Chain Monte Carlo for Unsupervised MRF Color Image SegmentationProceedings of Brithish Machine Vision Conference (BMVC) %I BMVA %P 37 - 46 %8 2004.09 %G eng %U http://www.bmva.org/bmvc/2004/papers/paper_223.pdf %0 Conference Paper %B A Képfeldolgozók és Alakfelismerők Társaságának konferenciája - KÉPAF 2004 %D 2004 %T Reversible Jump Markov Chain Monte Carlo for Unsupervised MRF Color Image Segmentation %B A Képfeldolgozók és Alakfelismerők Társaságának konferenciája - KÉPAF 2004 %C Miskolctapolca %P 144 - 151 %8 2004.01.28 %G eng %0 Conference Paper %B A Képfeldolgozók és Alakfelismerők Társaságának konferenciája - KÉPAF 2004 %D 2004 %T Számítógépes képfeldolgozás oktatása a Szegedi Tudományegyetemen %X

Az SZTE Informatikai Tanszékcsoportja által gondozott szakoktanterveiben 1993 óta szerepel a képfeldolgozás és alkalmazásainak oktatása. A kreditrendszer bevezetésével a Képfeldolgozás I. tárgy kötelező az ötéves képzésben részt vevő informatikus hallgatóknak. Ezen felül a választható szakirányok között szintén szerepel a Képfeldolgozás szakirány. A szakirányon belül különböző képpfeldolgozási területeket tárgyaló kurzusok épülnek egymásra. Az elméleti megalapozás mellett a képfeldolgozás alkalmazásaira is nagy hangsúlyt fektetünk. A kutatások illetve az orvosi alkalmazások fejlesztése során szerzett eredményeket a kötelező jellegű tárgyak mellett speciálkollégiumok keretében építjül be az otkatási anyagba. Számos hallgatónk választ a képfeldolgzás területéről témát a diplomamunkájához, dolgozataikkal rendszeresen és sikerrel szerepelnek az OTDK-n. Hallgatóink évente több hónapot tölthetnek külföldi partneregyetemeinken, ahol a kutató- és fejlesztőmunka mellett nálunk is elfogadott kurzusokat teljesíthetnek. A képfeldolgozás témakörön belül "ipari" projekt munkákban is egyre több hallgató vesz részt. A doktori programon belül is meghirdetünk képfeldolgozáshoz kapcsolódó kutatási irányokat. Az évente megrendezésre kerülő, 11-éves múltra visszatekintő Képfeldolgozó Nyári Iskolának (SSIP) eddig hatszor adott otthont Szeged. A rendszvénysorozat kiemelkedő fontosságú nemzetközi fórum hallgatóink és oktatóink számára is.

%B A Képfeldolgozók és Alakfelismerők Társaságának konferenciája - KÉPAF 2004 %I Neumann János Számítógép-tudományi Társaság %C Miskolc %P 191 - 196 %8 Jan 2004 %G eng %0 Book Section %B Proceedings of 11th Pacific Conference on Computer Graphics and Applications (PG) %D 2003 %T Non-Photorealistic Rendering and Content-Based Image Retrieval %B Proceedings of 11th Pacific Conference on Computer Graphics and Applications (PG) %I IEEE Computer Soc. Pr. %C New York %P 153 - 162 %8 2003/// %G eng %0 Book Section %B ICIP 2003: IEEE International Conference on Image Processing %D 2003 %T Unsupervised segmentation of color textured images using a multi-layer MRF model %X

Herein, we propose a novel multi-layer 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.

%B ICIP 2003: IEEE International Conference on Image Processing %I IEEE %P 961 - 964 %8 2003/// %G eng %0 Conference Paper %B IEEE - International Conference on Image Processing: ICIP %D 2002 %T Content-based image retrieval using stochastic paintbrush transformation %B IEEE - International Conference on Image Processing: ICIP %I IEEE Computer Society Press %C Aix-en-Provence %P 944 - 947 %8 Sep 2002 %G eng %0 Journal Article %J JOURNAL DE PHYSIQUE IV %D 2002 %T Markov random fields in image processing application to remote sensing and astrophysics %B JOURNAL DE PHYSIQUE IV %V 12 %P 117 - 136 %8 2002/// %@ 1155-4339 %G eng %N 1 %! J PHYS IV %0 Book Section %B Proceedings 16th International Conference on Pattern Recognition (ICPR 2002) %D 2002 %T Multicue MRF image segmentation: Combining texture and color features %X

Herein, we propose a new Markov random field (MRF) image segmentation model which aims at combining color and texture features. The model has a multi-layer 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. © 2002 IEEE.

%B Proceedings 16th International Conference on Pattern Recognition (ICPR 2002) %I IEEE Computer Society %P 660 - 663 %8 2002/// %G eng %0 Journal Article %J PATTERN RECOGNITION LETTERS %D 2001 %T Color image segmentation and parameter estimation in a markovian framework %X

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. © 2001 Elsevier Science B.V. All rights reserved.

%B PATTERN RECOGNITION LETTERS %V 22 %P 309 - 321 %8 2001/// %@ 0167-8655 %G eng %N 3-4 %! PATTERN RECOGN LETT %0 Book Section %B Proceedings of International Conference on Computer Analysis of Images and Patterns (CAIP) %D 2001 %T A Markov Random Field Image Segmentation Model Using Combined Color and Texture Features %B Proceedings of International Conference on Computer Analysis of Images and Patterns (CAIP) %I Springer Verlag %C Berlin; Heidelberg %P 547 - 554 %8 2001/// %G eng %0 Journal Article %J REAL-TIME IMAGING %D 2000 %T Image segmentation using Markov random field model in fully parallel cellular network architectures %B REAL-TIME IMAGING %V 6 %P 195 - 211 %8 2000/// %@ 1077-2014 %G eng %U http://www.sztaki.hu/~sziranyi/Papers/Sziranyi_MRF.pdf %N 3 %! REAL-TIME IMAGING %0 Generic %D 1999 %T Bayesian Color Image Segmentation Using Reversible Jump Markov Chain Monte Carlo %8 1999/// %G eng %0 Report %D 1999 %T Bayesian Color Image Segmentation Using Reversible Jump Markov Chain Monte Carlo %A Zoltan Kato %I ERCIM/CWI %C Amsterdam, The Netherlands %8 January 1999 %G eng %9 Research Report %0 Journal Article %J PATTERN RECOGNITION %D 1999 %T Unsupervised parallel image classification using Markovian models %X

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. © 1999 Pattern Recognition Society. Published by Elsevier Science Ltd. All rights reserved.

%B PATTERN RECOGNITION %V 32 %P 591 - 604 %8 1999/// %@ 0031-3203 %G eng %N 4 %! PATTERN RECOGN %0 Book Section %B Proceedings of Asian Conference on Computer Vision (ACCV) %D 1998 %T Motion Compensated Color Video Classification Using Markov Random Fields %B Proceedings of Asian Conference on Computer Vision (ACCV) %I Springer Verlag %C Berlin; Heidelberg %P 738 - 745 %8 1998/// %G eng %0 Conference Paper %D 1997 %T Color Image Classification and Parameter Estimation in a Markovian FrameworkProceedings of Workshop on 3D Computer Vision %P 75 - 79 %8 1997.05 %G eng %U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.128.1560 %0 Generic %D 1997 %T Image segmentation using Markov random field model in fully parallel cellular network architectures. %P - 17 %8 1997/// %G eng %0 Generic %D 1997 %T Markov Random Field Image Segmentation using Cellular Neural Network %8 1997/// %G eng %0 Generic %D 1997 %T Motion Compensated Color Image Classification and Parameter Estimation in a Markovian Framework %8 1997/// %G eng %U http://biblioteca.universia.net/html_bura/ficha/params/title/motion-compensated-color-image-classification-and-parameter-estimation-in-markovian/id/5664082.html %0 Conference Paper %B A Képfeldolgozók és Alakfelismerők Társaságának konferenciája - KÉPAF 1997 %D 1997 %T MRF based image segmentation with fully parallel cellular nonlinear networks %B A Képfeldolgozók és Alakfelismerők Társaságának konferenciája - KÉPAF 1997 %I Pannon Agrártudományi Egyetem Georgikon Mezőgazdaságtudományi Kar %C Keszthely %P 43 - 50 %8 Oct 1997 %G eng %0 Journal Article %J IMAGE AND VISION COMPUTING %D 1996 %T Bayesian image classification using Markov random fields %X

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 convexifies 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).

%B IMAGE AND VISION COMPUTING %V 14 %P 285 - 295 %8 1996/// %@ 0262-8856 %G eng %N 4 %! IMAGE VISION COMPUT %0 Book Section %B 1996 FOURTH IEEE INTERNATIONAL WORKSHOP ON CELLULAR NEURAL NETWORKS AND THEIR APPLICATIONS, PROCEEDINGS (CNNA-96) %D 1996 %T Cellular Neural Network in Markov Random Field Image Segmentation %B 1996 FOURTH IEEE INTERNATIONAL WORKSHOP ON CELLULAR NEURAL NETWORKS AND THEIR APPLICATIONS, PROCEEDINGS (CNNA-96) %I Wiley - IEEE Press %C New York %P 139 - 144 %8 1996/// %G eng %0 Journal Article %J GRAPHICAL MODELS AND IMAGE PROCESSING %D 1996 %T A Hierarchical Markov Random Field Model and Multitemperature Annealing for Parallel Image Classification %X

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 multitemperature 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 generalization of the annealing theorem of S. Geman and D. Geman (IEEE Trans. Pattern Anal. Mach. Intell. 6, 1984, 721-741). © 1996 Academic Press, Inc.

%B GRAPHICAL MODELS AND IMAGE PROCESSING %V 58 %P 18 - 37 %8 1996/// %@ 1077-3169 %G eng %N 1 %! GRAPH MODEL IM PROC %0 Journal Article %J IEEE TRANSACTIONS ON IMAGE PROCESSING %D 1995 %T DPA: a deterministic approach to the MAP problem %X 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. %B IEEE TRANSACTIONS ON IMAGE PROCESSING %V 4 %P 1312 - 1314 %8 1995/// %@ 1057-7149 %G eng %N 9 %! IEEE T IMAGE PROCESS %0 Book Section %B ICASSP-95 %D 1995 %T Unsupervised adaptive image segmentation %X 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. %B ICASSP-95 %I IEEE %C Piscataway %P 2399 - 2402 %8 1995/// %G eng %0 Book Section %B Proceedings of the 5th International Conference on Computer Vision %D 1995 %T Unsupervised parallel image classification using a hierarchical Markovian model %X 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...). 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). %B Proceedings of the 5th International Conference on Computer Vision %I IEEE %C Piscataway %P 169 - 174 %8 1995/// %G eng %0 Thesis %D 1994 %T Multi-scale Markovian Modelisation in Computer Vision with Applications to SPOT Image Segmentation : Modélisations markoviennes multirésolutions en vision par ordinateur. Application ŕ la segmentation d'images SPOT %8 1994 %G eng %0 Book Section %B Proceedings of the 12th IAPR International Conference on Pattern Recognition %D 1994 %T Multi-Temperature Annealing: A New Approach for the Energy-Minimization of Hierarchical Markov Random Field Models %B Proceedings of the 12th IAPR International Conference on Pattern Recognition %I IEEE %C Los Alamitos %P 520 - 522 %8 1994/// %G eng %0 Generic %D 1994 %T Segmentation hiérarchique d'images sur CM200 (Hierarchical Image Segmentation on the CM200) %8 1994/// %G eng %0 Generic %D 1994 %T Segmentation multirésolution d'images sur SUN version 1 du 26.05.1994 (Multiresolution Image Segmentation on SUN version 1 of 26.05.1994) %8 1994/// %G eng %U http://www.app.asso.fr/en/ %0 Book Section %B Maximum Entropy and Bayesian Methods %D 1993 %T Bayesian Image Classification Using Markov Random Fields %B Maximum Entropy and Bayesian Methods %I Kluwer Academic Publishers %C Dordrecht; Boston; London %P 375 - 382 %8 1993/// %G eng %0 Generic %D 1993 %T Extraction d'information dans les images SPOT %8 1993/// %G eng %0 Generic %D 1993 %T A Hierarchical Markov Random Field Model and Multi-Temperature Annealing for Parallel Image Classification %8 1993/// %G eng %U http://hal.inria.fr/inria-00074736/ %0 Conference Paper %B International Workshop on Image and Multidimensional Digital Signal Processing (IMDSP) %D 1993 %T A Hierarchical Markov Random Field Model for Image Classification %B International Workshop on Image and Multidimensional Digital Signal Processing (IMDSP) %I IEEE Computer Soc. Pr. %8 Sep 1993 %G eng %0 Book Section %B Fourth International Conference on Computer Vision, ICCV 1993, Berlin, Germany, 11-14 May, 1993, Proceedings %D 1993 %T Multiscale Markov random field models for parallel image classification %X 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. %B Fourth International Conference on Computer Vision, ICCV 1993, Berlin, Germany, 11-14 May, 1993, Proceedings %I IEEE %C Los Alamitos %P 253 - 257 %8 1993/// %G eng %0 Book Section %B ICASSP-93 %D 1993 %T Parallel image classification using multiscale Markov random fields %X 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. %B ICASSP-93 %I IEEE %C New York %P 137 - 140 %8 1993/// %G eng %0 Generic %D 1992 %T Image Classification Using Markov Random Fields with Two New Relaxation Methods %8 1992/// %G eng %U http://hal.inria.fr/docs/00/07/49/54/PDF/RR-1606.pdf %0 Conference Paper %B International Conference on Acoustics, Speech and Signal Processing (ICASSP) %D 1992 %T Satellite Image Classification Using a Modified Metropolis Dynamics %B International Conference on Acoustics, Speech and Signal Processing (ICASSP) %I IEEE Computer Soc. Pr. %P 573 - 576 %8 Mar 1992 %G eng