MRI brain tumour segmentation based on fish chaining transition optimization algorithm

In image processing, images are aligned into regions of similar features by means of segmentation. MRI is an imaging modality which acts as an input source for the researchers and the doctors to identify the tumour in the brain. More segmentation algorithms have been proposed by researchers for tumour image segmentation. In this proposed work, a bio-inspired technique, swarm intelligence based fish characteristics approach is used to segment the tumour region. In this automatic MRI brain tumour segmentation, initially, the obtained MRI brain models are partitioned into equal symmetric halves. Then, the skull is marked and the geometric shape of the ellipse is personalized over the skull outer region. Finally, the proposed Fish Chaining and Transition Optimization Algorithm (FCTOA) is used to detect the tumour. The proposed FCTOA is compared with the various existing algorithms such as FCM, SOM and PSO. The experimental results of the FCTOA obtain better performance and show 0.98 Sensitivity, 0.92 Specificity, 0.89 Dice similarity coefficient, 0.83 Positive predictive value, 7.4774 mm Hausdorff distance and 0.8018 Jaccard coefficient. Key words: MRI, segmentation, brain tumour, optimization algorithm, medical image