Segmentation and Analysis of Corpus Callosum in Autistic MR Brain Images Using Reaction Diffusion Level Sets

In this work, corpus callosum of control and autistic MR brain images are segmented using reaction diffusion level set method. The level set function is initialized manually. The evolution of level set contour is regularized by reaction diffusion method. Two different stopping functions are analysed to converge the level set function. The corpus callosum segmented using level set method is validated with the ground truth. Results show that reaction diffusion level set method is able to segment the corpus callosum from other brain segments. Also, the corpus callosum segmented using reaction diffusion level set gives high value of similarity with the ground truth. The stopping function based on area inside the region of interest gives high value of similarity measures for segmentation of corpus callosum with less number of iterations. The area calculated from the reaction diffusion based level set using the above stopping criterion and ground truth gives high correlation (R = 0.99) for both control and autism images. As the proposed method of segmentation and analysis. seems to be clinically significant, this can be used for the mass screening of autism like brain disorders.