Breast Cancer Pathological Image Auto-Classification using Weighted GLCM-SVM

At present, when radiologists screen the breast cancer, they always miss about 10-30% tumors for pathological edge blurring and asthenopia after the diagnosis for a long time. For this reason, setting up the Computer Aided Diagnosis (CAD) system is of great clinical application value. This thesis proposes the method of weighted GLCM to extract the texture features of mammogram and classify abnormal regions according to the evaluation of the performance of each individual by using support vector machines (SVM). The experiment shows our system is effective for breast cancer pathological image.