The application of statistic image analysis for classification of breast cancer based on mammograms

Abstract

Cancer is a disease characterized by the ability of abnormal cells to grow uncontrollably. In the medical field, detection of breast cancer is done using a mammogram. Examination of the mammogram image is still done manually by the doctor / radiologist, so it is necessary to use technology as supporting information. In this research, mammogram image classification based on gray-level co-occurrence (GLCM) matrix and gray-tone difference matrix (GTDM) has been done with backpropagation method. The stages of the mammogram image classification process include the process of image acquisition, pre-processing, feature extraction with GLCM and GTDM and classification using backpropagation. The preprocessing process carried out is gray-scalling, contrast enhancement, image segmentation with Otsu thresholding, edge detection process, and image thickening process with widening morphology method. The highest performance results for accuracy are 85% and precision is 85.7%. This result was obtained when using the GLCM and GTDM feature extraction methods.