Performance analysis of cad system in detection of abnormalities in mammography images

Abstract

Cancer is the major cause for death of human beings in the world. There are many types in cancer, out of which breast cancer is the most common cause for death of women. Medical image processing is an emerging area, which helps the medical practitioners to diagnose the cancer and its surgery. Mammography is one of the existing imagi ng techniques used for detection of breast cancer. Masses and calcifications, observed from the mammogram are the indicators of cancer. The decision taken about cancer is probably incorrect, if noise is present in the mammographic image. To filter these noises various algorithms have been developed, but they may cause some unpleasant visual artifacts. Moreover, the existing diagnostic techniques are inadequate to examine and interpret the data of breast cancer image. Also the detection accuracy of breast cancer is poor and there are difficulties in handling large dataset. Therefore, radiologists want a sophisticated tool to categorize the cancer. To develop such a sophisticated tool, large number of medical images with their diagnosis report has been considered. In this research work, a Computer Aided Diagnosis (CAD) system is developed for detection and classification of abnormalities in the medical images. A high boost gaussian filtering technique is proposed to suppress the noise and enhance the mammographic image quality, so as to diagnose the cancer appropriately. newline