Analysis and Improvement in Image Segmentation for CT Images

Abstract

Image segmentation in medical images is a very complex task. Currently, brain diseases are detected by imaging only after the appearance of neurological or nervous system symptoms. Manual segmentation is a long and painful task. This method is not reliable and error sensitive. The need for correct segmentation of the ailment is very important for proper medications as any delay or wrong diagnosis may become fatal to the patient. Many methods have been developed to segment tumor. In this work, we have used Computed Tomography (CT) images for segmentation of abnormal portion. Two techniques, modified region based active contour and hybrid level set method, have been used in this thesis work. In the former method, the number of iterations reduced as well as the time consumption drastically reduced. The latter method, overcomes the problem of under segmentation as well as over segmentation.