Detection of Microcalcification with Top-Hat Transform and the Gibbs Random Fields
Published in 37th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC), 2015
This paper proposes a multi-stage computer-aided detection (CAD) pipeline for microcalcifications (MCs) in digital mammograms. Breast cancer is among the leading causes of death in women, and MCs — tiny bright spots as small as 0.1mm — are early indicators of malignancy. The method begins with fuzzy C-means clustering for gross breast-region segmentation, followed by top-hat transform to localize the region of interest, and a watershed transform for precise isolation. Gibbs random field clique patterns are then used to detect MC candidates via adaptive thresholding. The algorithm achieves an overall detection rate of 94.4%, sensitivity of 93.7%, and detection accuracy of 88.2%, with a false-negative rate of only 5.6%.
Recommended citation: Bharadwaj, A. S. & Celenk, M. (2015). "Detection of Microcalcification with Top-Hat Transform and the Gibbs Random Fields." 37th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC). Milan, Italy. p. 6382.
Download Paper