Detection of cancer tumors in mammography images using support vector machine and mixed gravitational search algorithm

Abstract

In this paper, support vector machine (SVM) and mixed gravitational search algorithm (MGSA) are utilized to detect the breast cancer tumors in mammography images. Sech template matching method is used to segment images and extract the regions of interest (ROIs). Gray-level co-occurrence matrix (GLCM) is used to extract features. The mixed GSA is used for optimization of the classifier parameters and selecting salient features. The main goal of using MGSA-SVM is to decrease the number of features and to improve the SVM classification accuracy. Finally, the selected features and the tuned SVM classifier are used for detecting tumors. The experimental results show that the proposed method is able to optimize both feature selection and the SVM parameters for the breast cancer tumor detection.