Automatic detection of clustered microcalcifications in digitized mammogram films

The existence of clustered microcalcifications is one of the important earl y signs of breast cancer. This paper presents an image processing procedure for the automatic detection of clustered microcalcifications in digitized mammograms. in particular, a sensitivity range of around one false positive per image is targeted The proposed method consists of two main steps. Firs t, possible microcalcification pixels in the mammograms are segmented out u sing wavelet features or both wavelet features and gray level statistical f eatures, and labeled into potential individual microcalcification objects b y their spatial connectivity. Second, individual microcalcifications are de tected by using the structure features extracted from the potential microca lcification objects. The classifiers used in these two steps are feedforwar d neutral networks. The method is applied to a database of 40 mammograms (N ijmegen database) containing 105 clusters of microcalcifications. A free re sponse operating characteristics curve is used to evaluate the performance. Results show that the proposed procedure gives quite satisfactory detectio n performance. In particular, a 93% mean true positive detection rate is ac hieved at the price of one false positive per image when both wavelet featu res and gray level statistical features are used in the first step. (C) 199 9 SPIE and IS&T. [S1017-9909(99)00701-1].