STUDY OF FCM WITH VARIOUS OPTIMIZATION ALGORITHMS FOR IMAGE SEGMENTATION

Abstract

With computer-aided diagnostics, medical pictures are acquired utilizing electronic instruments such as CT scanners and MRI machines. Typically, computed tomography (CT)/magnetic resonance imaging (MRI) pictures obtained are restricted spatial resolution, low contrast, noise, and nonuniform variation because of environmental impacts. As a result, the peculiarities of the objects are fuzzy and twisted, and their meanings are not precise. Fuzzy sets and fuzzy logic are best suited for dealing with ambiguity and uncertainty. For segmentation, the fuzzy clustering approach has been widely employed and assorted photos during the previous decade. Fuzzy clustering is a significant challenge that is being actively researched in various real-world applications. One of the algorithms is the fuzzy c-means (FCM). since it is the most prevalent fuzzy clustering approach, clear and simple to implement. FCM on the other hand can stuck in local optima and is sensitive to initialization. This study compares seven fuzzy-clustered image segmentation methods utilized in CT scan and MRI brain image segments. The algorithms were examined using CT scan/MRI brain images in this investigation. The methods were statistically assessed in the research using two validity metrics, partition coefficient and partition entropy.