SEGMENTATION OF MEDICAL IMAGES THROUGH COMPETITIVE LEARNING

In image analysis applications, segmentation of gray-level images into meaningful regions is an important low-level processing step. Various approaches to segmentation investigated in the literature, in general , use either local information of gray-level values of pixels (region growing based methods, for example) or the global information (histogr am thresholding based methods, for example). Application of these appr oaches for segmenting medical images often does not provide satisfacto ry results. Medical images are usually characterized by low local cont rast and noisy or faded features causing unacceptable performance of l ocal information based segmentation methods. In addition, because of a large amount of structural information found in medical images, globa l information based segmentation methods yield inadequate results in r egion extraction. We present a novel approach to image segmentation th at combines local contrast as well as global gray-level distribution i nformation. The presented method adaptively learns useful features and regions through the use of a normalized contrast function as a measur e of local information and a competitive learning based method to upda te region segmentation incorporating global information about the gray -level distribution of the image. In this paper, we present the framew ork of such a self organizing feature map, and show the results on sim ulated as well as real medical images.