ABSTRACT
OBJECTIVE: Breast cancer represents themost frequently diagnosed cancer in women. In order to reduce mortality, early detection of breast cancer is important, because diagnosis is more likely to be successful in the early stages of the disease. On the average, the reader's sensitivity can be increased by 10%with the assistance of computer-aided diagnosis (CAD) system. This paper presents a CAD system for the automatic detection of clustered micro-calcifications in digitized mammograms. METHODS: The proposed system consists of three main steps. First, breast region is segmented from original mammogram using contrast property of grey level co-occurrence matrix(GLCM). Second, potential micro-calcification pixels in the mammograms are detected by foveal method. Third, in order to reduce false-positive rate, individual micro-calcifications are detected by a set of 8 features extracted from the potential individual micro-calcification objects. RESULTS: In the result, Specificity and sensitivity are used to evaluate the detection performance of micro-calcifications.(sensitivity : 93.1%, specificity : 87.5%). CONCLUSION: This study could be a useful method for diagnosis of breast cancer as a CAD system.