Segmentation of breast images to find the lesions using the decreased neighborhood graph algorithm

Classification of soft tissues is often joined with uncertainty and ultimate areas border might be hardly measured in segmentation. Basic techniques of edge detection can be used to determine the boundary edge, but because of noise and gray levels steep changes in medical images, it is difficult to achieve the edge of a lesion in image with reasonable and precise segmentation. In this paper a graph based algorithm is presented to extract alike region in image, this approach helps finding similar lesions and accurate segmentation.