Image Segmentation Techniques

1. Goals – build and test Matlab implementations of:

   • Active Contour Analysis (Szeliski 5.1)
   • Mean shift (Szeliski 5.3)
   • Expectation Maximization [E-M] method – mixture of Gaussians

2. Sources

   • Szeliski
   • Expectation Maximization
     • http://courses.cs.washington.edu/courses/cse576/12sp/notes/remote.pdf

3. Report

   • Give mathematical and algorithmic theory for each method
   • Describe Matlab implementations and how to run the Matlab code
   • Test methods with gray-scale and color images
     • Vary parameters
   • Give well-commented code
   • Sum up observations on the pro’s and cons of the methods

4. Due date – Wednesday 18 December (during Final’s Week)