Segmentation of microcalcifications in mammograms based on multi-resolution region growth and image difference

Abstract

Effective segmentation of microcalcifications in mammograms is crucial for the quantification of morphologic properties by features incorporated in computer-aided diagnosis schemes. A multi-resolution region growth method based on edge feature is proposed in this paper, in which edge feature vectors are used to obtain complete microcalcifications. Then a microcalcification segmentation method based on the multi-resolution region growth method and image difference is presented. By using this method, under the precondition of high detecting rate, complete microcalcifications can be segmented. What is more, the shape and distribution can be obtained accurately, and false positive can be decreased, which can better support subsequent feature extraction and microcalcification classification. Lots of experiment results prove that with the method, the detecting rate and false positive of microcalcifications can reach higher than 94% and lower than 2.00, ahead of that by domestic and foreign well-known algorithms.