SSIP'97 Projects

Project 1 Project 2 Project 3 Project 4
Project 5 Project 6 Project 7 Project 8
Project 9 Project 10 Project x1 Project x2

Project 1 -

Label the group photo- identify faces and label them. Try to generate atlas of photo from the group photo.

Remarks: Difficulty medium

Project 2 - (Team D) - (PPT presentation)

Extraction of the brain (or part of the brain) from a 3-D MR data set by region growing/ edge detection.

Input: a 3-D data set, for example MR brain.

Operation: given a seed point generated manually, and a criteria defining 'edge' for example voxel must have values between two limits, grow the region in 3-D to extract the selected object.


Operation: By convolution with 3x3x3 operators, to identify the 3-D edge and hence extract the object or part of the object.

Output: data set viewable with 3Dviewnix

Coding: in C or under PIP

Remarks: Difficulty easy. Manual editing could be added. In competition with 2.

Project 3 -

Block world line drawing generated from range image input of scene.

Input: 2-D Image of 3-d block world as a range image (pixel value function of distance)

Operation: Identification of lines and corners, linking of this to extract model of scene.

Output: Line drawing of scene. Rotation of object to create new scene.

Coding: C

Remarks: Difficulty hard.

Project 4 - (Team C) - (HTML presentation)

Extraction of 'skeleton' from 3-D objects.

Project 5 -

Registration of 2-D images using critical point matching.

Input: CT/MR images

Operation: extraction of contour, identification of (geometrical) critical points (maximum gradient, 2nd differential etc of contour), computation of transformation to register two such images (e.g. by SVD)

Output: Transformed image

Coding: as desired (but not GUI)

Remarks: Difficulty medium

Project 6 -

Motion correction

Given a set of images in time which are similar but not identical, derive a method for identifying the motion that has occurred (in 2d) between different images (shift and rotation) such that they can be adjusted and corrected.

Input: Starting from image sequence, identify features for example with skeleton, or determin regions of change by subtraction etc. Find significant changes

Output: image marked with regions of change

Coding: as desired

Remarks: Difficulty medium

Project 7 - (István Nagy) - (PPT presentation)

Create a cine display of a sequence of objects using Borland C++ builder, with control over speed etc

Remarks: Difficulty: Who knows (problem is documentation)

Project 8 -

Generate a compute based logo for IPMI, using bend of the river at Visegrad and the number 16.

Ouput: for exmaple a Java applet

Remarks: Difficulty: requires artictic ability

Project 9 - (Team B) - (PPT presentation)

Texture analysis using wavelet transform.

Input: MR (or radiology) 2-D image

Operation: Application of wavelet transform (see Numerical recipes- which needs to be downloaded) to give regional texture information, or other code available

Output: Texture map.

Coding: as desired (but not development of GUI)

Remarks: Difficulty medium. Testing using simulated data also required.

Project 10 - (Team A) - (PPT presentation)

Matched filter v. observer

Project x1 -

2-D edge detection using cost minimization/ snakes.

Input: 2-D nuclear medicine data or vocal folds images.

Operation: Define a transform, for example polar, a cost function, for example circumference and gradient. Minimize path in transformed data by cost minimization.


Fit a snake for example using Greedy algorithm


Sobel operator to give edge strength and direction maps.

Operation: To find an algorithm to link the points identified on these map to give continuous enclosing contours.

Output: Image with contour. Algorithm to identify organ, for example left ventricle of heart, without manual intervention.

Coding: In C or under PIP but in form which could be used in package

Remarks: Difficulty medium. Problem is robustness

Project x2 - (Team E) - (PPT presentation)

Match of fragment of coastline to map

Starting off with a segment of coastline from a map, of different scale and noise properties extracted from a (much) larger segment, perform a best fit to identify the section of coastline. One method that could be used would be by correlation of a chain code representation.

Creation of chain code or equivalent from map segment is part of project

Input: 2d map

Output: contour with match

Coding: as desired

Difficulty: quite easy