Graph Cuts based Image Segmentation using Fuzzy Rule Based System

This work deals with segmentation of the gray scale, color and texture images using graph cuts. From an input image, a graph is constructed using intensity, color and texture profiles of the image simultaneously (i.e., intensity and texture for gray scale images and color and texture for color images). Based on the nature of image, a fuzzy rule based system is designed to find the weight that should be given to a specific image feature during graph development. The graph obtained from the fuzzy rule based weighted average of different image features is further used in normalized graph cuts framework. The graph is iteratively bipartitioned through the normalized graph cuts algorithm to get optimum partitions resulting in segmented image. The Berkeley segmentation database is used to test our algorithm and the segmentation results are evaluated through probabilistic rand index, global consistency error, sensitivity, positive predictive value and Dice similarity coefficient. It is shown that the presented segmentation method provides effective results for most types of images.