Segmentation of pectoral muscle and detection of masses in mammographic images

A segmentation method is proposed where the pectoral muscle in mammograms is segmented and then the mass is detected. First two anatomical features of the pectoral muscle, homogeneous texture and high intensity deviation are employed to identify the initial pectoral muscle edge. The maximum searching algorithm is used to obtain the initial edge. However the obtained initial edge in many cases is not refined. Then the Kalman filter is used to refine the ragged initial edge. Once the refined edge is obtained the mass is detected using gamma distribution and then the contour is detected using region based active contour model to obtain the segmented result.