Graph cut segmentation technique for MRI brain tumor extraction

In this paper, we present a graph cut application dealing with MRI brain image segmentation. We here propose another emerging approach of region segmentation based on graph cut which supports on the eigenspace characteristics and the perceptual grouping properties to classify brain tumoral tissue. Image segmentation is considered as solving the partitioning clustering problem by extracting the global impression of image. In the aim of providing visual and quantitative information for the diagnosis help in brain diseases, tumor features observed in image sequences must be extracted efficiently. We lastly extend this approach to perform volume segmentation by matching 2D contours set. This 3D representation provides a precise quantitative measure for following up the tumor brain evolution in the case of patients under pharmaceutical treatments.