HYBRID WCA–SCA AND MODIFIED FRFCM TECHNIQUE FOR ENHANCEMENT AND SEGMENTATION OF BRAIN TUMOR FROM MAGNETIC RESONANCE IMAGES

Abstract

Enhancement of image plays an important and vital role in preprocessing the magnetic resonance images (MRI). At the same time, image segmentation techniques are also essential to detect and remove the noise to enhance the quality of MRI to detect the infected regions of the brain tumor. This paper presents a novel image enhancement technique for preprocessing of brain tumor MRI by hybridizing the Water Cycle Algorithm (WCA) and Sine Cosine Algorithm (SCA). The WCA is based on the process of water cycle in rivers and streams flow in the ocean whereas the SCA follows the cyclic form of sine and cosine trigonometric functions, which permits a search agent to be transposed around the desired solution. In fact, the Fuzzy [Formula: see text] means-based segmentation algorithms have proved their ability in automatic detection of the tumor and help doctors and radiologist to diagnose the type of tumor from the MRI, but, some of the FCM-based algorithms fail to remove the required amount of noise from the MRI which restrict doctors to have better segmentation accuracy. A modified fast and robust FCM (MFRFCM) segmentation technique has been proposed to sharpen and remove noise from MRI to detect the brain tumor to have improved accuracy. In this research work, Dataset-255 is considered from the Harvard medical school. The results from the proposed hybrid WCA-SCA technique are compared with WCA, SCA and comparison results are presented. The hybrid WCA+SCA image enhancement technique attains an accuracy of 99.25% for benign tumor and 98.52% for malignant tumor. Further, the results of modified Fast and Robust FCM (MFRFCM) segmentation results are compared with the conventional FCM-based segmentation algorithms. It is observed that the proposed hybrid WCA-SCA image enhancement technique and modified FRFCM Segmentation outperform in terms of computational time and performance accuracy in contrast to the other algorithms.