A model based algorithm to segment the pectoral muscle in mammograms

In this paper, a model-based algorithm is described to segment the pectoral muscle in the mammograms. After the mammogram has been preprocessed, a set of ROI with different sizes is applied on it, in each of which, an iterative thresholding technique is used to gain an optimal threshold correspondingly. All the thresholds compose the threshold curve and e curve, which the optimal threshold of the mammogram could be extracted by. Finally, a twice line fitting and polygon approaching technique is carried out to refine and approach the edge curve of the pectoral muscle, which has been partly segmented by the thresholding. About 60 mammograms have been used to evaluate the proposed algorithm with satisfactory result reached.