Efficient implementation for classifying and segmenting of computed tomography brain tumour images using modified region growing with lion algorithm. (2017)

• Record Type: Journal Article
• Authors: Ramakrishnan, Thiagarajan; Sankaragomathi, Balasubramanian
• Is Part Of: International journal of biomedical engineering and technology. Volume 23:Number 2/4(2017)
• Page Start: 159
• Page End: 179
• Publication Date: 2017
• Subjects: computed tomography -- CT images -- feature extraction -- support vector machines -- SVM -- SMO -- sequential minimal optimisation -- quadratic programming -- least squares -- classification -- modified region growing -- threshold optimisation -- lion algorithm -- image segmentation -- classification -- brain tumour images -- medical images -- brain tumours; Biomedical engineering -- Periodicals; 610.28
• Languages: English
• ISSNs: 1752-6418

Abstract

The study of Computed Tomography (CT) images considered in image segmentation is very important as it plays a vital part in identifying the different kinds of tumour. The classification of the tumour and the non-tumour images is followed by the segmentation of tumour region in CT images which are done by the proposed methodology. The process of classifying is done by Support Vector Machine (SVM) with linear kernel as well as Sequential Minimal Optimisation (SMO). After the classification process, segmentation is performed by the Modified Region Growing (MRG) with a threshold optimisation using Lion Algorithm (LA). In comparison, the sensitivity value of the proposed approach is appreciably higher as 91.52% than that of the existing method of MRG-Grey Wolf Optimisation (MRG-GWO). The comparative analysis in terms of sensitivity, specificity and accuracy is done for the proposed as well as for the existing techniques.