ABSTRACT:- Breast cancer seems to be the world’s-largest cause of the death among women. Currently, screening mammography is the best scientific radiological strategy for early diagnosis of breast cancer. However, Segmentation for region of Interest (ROI) is the first and critical step in evaluating digital mammogram images for breast cancer. For this purpose, the preprocessing of mammogram images is very important in the breast cancer diagnosis process as it decreases the number of mammogram images. This paper suggests different strategies for solving noise removal problems and by using a thresholding method, separating the background region from the breast profile section. Finding a precise, robust and effective ROI segmentation technique in digital mammography analysis is still a challenging issue. In this paper we proposed an efficient morphological approach for segmenting interest regions from mammogram images. For research purposes, digital mammograms are taken from the mini MIAS (Mammography Image Analysis Society) database and the findings are collected in MATLAB. These results indicate that the proposed approach is efficient and reliable.

Keywords: Noise removal; pectoral muscle extraction; MATLAB; ROI segmentation