Precise Identification and Segmentation of Brain Tumour in MR Brain Images Using Salp Swarm Optimized K-Means Clustering Technique

Abstract

Brain tumour delineation is a challenging task from raw magnetic resonance images. To accurately delineate the different parts of tumours is the main aim of dissection process. Among the most common types of cerebral tumour, glioma that arises from glial cells. According to the World Health Organisation (WHO), tumour behaviours and microscopic images can be used to classify gliomas into four different levels. The popular imaging techniques used prior to and following surgical treatment is magnetic resonance imaging (MRI), which aims to provide vital details for the therapeutic plan. For effective tumour delineation from brain MRI, a novel combination of K-means and Salp Swarm Optimization (SSO) Algorithm is proposed. K-means clustering method groups the most similar pixels in to a single cluster. Salp Swarm Optimization Algorithm is one of the nature-inspired metaheuristic optimization algorithms based on the social and foraging behaviour of salps. In biomedical signal processing and control systems, SSO is used to tackle large-scale optimization problems. The proposed methodology's efficiency is validated through testing on various BraTS challenge datasets. The attained average computational time, MSE, PSNR, TC and DS are 16.9 Sec, 0.3787, 52.47 dB, 74.86 % and 83.44 %, respectively.