A GLCM Algorithm for Optimal Features of Mammographic Images for Detection of Breast Cancer

Abstract

Mammography is the golden standard for imaging and diagnosis of breast cancer in early stages. However, it is difficult for radiologists to interpret results from mammograms, particularly mammogram images have low contrast and the image quality related to technician and devices. In this paper, optimal features from a Gray Level Co-occurrence Matrix (GLCM) algorithm are applied to the mammography images for increseasing the accuracy of breast cancer detection. In particular, two datasets of the mammography images are filtered and segmented to produce Glandular Tissue Region (GTR) which contain significant features. Therefore, we just choose 4 optimal features of 10 ones using the GLCM algorithm through statistic evaluation. The results show that the optimal selected features have significant impact to produce breast cancer detection with high performance using a SVM classifier.