Medical Image Segmentation with Shape Priors
Master Semester Project
Image segmentation is a fundamental problem in computer vision and image processing. Segmentation models basically look for edges or/and homogeneous intensity regions to find meaningful objects lying in images. However the segmentation of structures of interest can not be successfully done in the presence of occlusions or strongly cluttered background. In order to deal with this kind of problems, we propose to develop a constrained parametric active contour model which incorporates prior information about the shape to segment. This prior shape will be derived by using pattern recognition techniques on the training shape. In this project the student will implement in Java a parametric active contour and study the impact of different shape priors and machine learning techniques.
- Ricard Delgado-Gonzalo, email@example.com, 021 693 51 43, BM 4.141
- Michael Unser, firstname.lastname@example.org, 021 693 51 75, BM 4.136