Brain tumor segmentation using normalized graph cuts

Abstract

Normalized graph cut algorithm is an efficient method where the technique of graph theory is adopted and in which the images are taken in the form of weighted graph in order to segment the images. This paper comprises of the fundamental concept of Normalized graph cut algorithm and its application towards the segmentation of Brain tumor. Identifying defects such as tumors is a very challenging because differentiating the components is difficult in a complex structure like a human brain. The diagnosis becomes even more complex because the tumor, blood clots and some part of the brain tissues appear as the same Brain tumor is generally detected and analyzed through a comprehensive analysis of the Magnetic Resonance Images of the brain. This technique gives a second opinion regarding the presence or absence of the brain tumor. This paper performs the study of Normalized graph cut algorithm and shows its efficiency in detecting tumors and compares it with other commonly used algorithms.