Classification of normal and cancerous mammogram images based on texture features using the Support Vector Machine (SVM) method

Abstract

Breast cancer is the leading cause of death in older women, with more than one million women worldwide dying from this disease yearly. Mammography is a specialized radiological examination that uses low-dose X-rays to detect breast abnormalities, even before visible symptoms such as palpable lumps appear. This study aims to develop an effective mammogram image classification model using the SVM (Support Vector Machine) method with texture feature extraction analysis in histograms and GLCM (Gray-Level Co-Occurrence Matrix). The research involved 20 normal and 20 cancer images, starting with mammogram image preprocessing, then texture feature extraction using histograms and GLCM, and ending with data classification using the SVM method. Test results showed that SVM could classify images with an accuracy of 67.5%, a sensitivity of 33.3%, and a specificity of 70%. Therefore, this research could be a foundation for further developments to enhance mammogram image classification accuracy.