A novel intuitionistic fuzzy co-clustering algorithm for brain images

Abstract

Segmentation of Magnetic Resonance Imaging(MRI) Brain images is a very important step in detection of brain tumor. This process is hand operated in labs which is rather an enervating and a long drawn out task and the resultant data so obtained has high degree of erroneity and inconsistency. Hence, automated segmentation systems have become the need of the hour. This paper presents a novel technique to automatically detect brain tumor. Also, segmentation of three main brain tissues is carried out namely white matter, gray matter and cerebrospinal fluid from real time Magnetic Resonance Imaging(MRI) images. The Intuitionistic Fuzzy Set theory is incorporated as it is more suitable for handling uncertainty as compared to fuzzy sets theory. The algorithm is based on the Co-Clustering approach as it offers the advantage of assigning membership functions to both object as well as features. The parameters in the IFCC algorithm are optimized using Particle Swarm Optimization(PSO). The performance of the algorithm is evaluated on the basis of quantitative measures such as match score, accuracy score, dice score and Jaccard’s similarity coefficient.