Welcome to the Intelligent Brain Segmentation Project Page.

A mixture of pattern recognition and expert system techniques are used to segment human brain images into tissue types of interest. The tissue types of interest are gray matter, white matter, cerebro-spinal fluid, tumor, edema and necrosis. The goal is to do each segmenation without human interaction.
A Magnetic Resonance Image slice of the brain. It has been segmented into tissue regions by a knowledge guided clustering system developed in the ISL. Red is white matter, orange is gray matter, yellow is csf, green edema and gray denotes tumor.

Some recent papers on knowledge guided image segmentation

1995 IJCAI Conference (html)

12th IAPR International Conference on Pattern Recognition, 1994 (html)

IEEE Eng. in Medicine & Biology v.13 no.5 pp.730-742 1994 (html)

  • Back to the Intelligent Systems Lab home page.