An introduction to a stochastic based color image segmentation method

Mr. Zhong Li

Computer Science Dept., UNC-Charlotte

Friday, September 21, 2007

3:00 p.m. - 4:00 p.m.

Woodward Hall, Room 106

Complete Description:

The presented method focuses on handle the large diversity of color in natural images. The method treats the segmentation procedure as a pattern recognition procedure in which the final segmentation is produced by the evolution and split / merge operation driven by the Gaussian mixture model. The method is adaptive to the individual image and no training process is needed. We also extend this method to 3D medical image and showing good result. Bio: Mr. Zhong Li is a Phd student under the guidance of Dr. Jianping Fan. More information about the research group can be found at this webpage.