A comparison of clustered microcalcifications automated detection methods in digital mammogram

This paper presents the comparison of three automated methods for an early detection of breast cancer. It specifically detects clusters of microcalcifications (MCCs), which are associated with a high probability of malignancy. The proposed methods are based on several image processing concepts, such as morphological processing, fractal analysis, adaptive wavelet transform, local maxima detection and high-order statistics (HOS) tests. We apply these methods on a set of mammograms (MIAS database) to test their efficiency and computation time. It shows that the HOS test proved to be the most efficient, and give reliable results for every mammogram tested.