An Hybrid Approach for Identification of Breast Cancer using Mammogram Images

Abstract

Breast Cancer (BC) is the first among the cancer deaths in women all over the world. Mammography is broadly perceived as the best imaging methodology for the early location of BC. Mammography examination reduced the BC death in spite of increasing number of noticed malignancies during the last decade. Although it is the best reliable method for early location, it has several limitations. One essential viewpoint is that the exactness rate tends to diminishing when doctors' examined high volume of mammograms. This work mainly concentrates on identifying regions containing small clusters of micro calcifications to categorize the tissue as being regular or irregular. Potentially cancerous tissue is distinguished from normal tissue by analyzing features of a given region within a mammogram. Therefore, feature extraction and saliency play an important role in cancer detection.