Abstract
In this report, a novel technique is proposed for computer-aided automatic extraction of microcalcifications in a digital mammogram. First, the microcalcifications are detected by morphological filtering, followed by entropy-based thresholding. Next, the microcalcifications are segmented by computing regional watershed. The proposed automatic technique is designed to serve as a visual aid to radiologists. Its efficacy is demonstrated through experimental results.
Full Text
The Full Text of this article is available as a PDF (1.2 MB).
References
- 1.Chang Y-H, Zheng B, Good WF, et al. Identification of clustered microcalcifications on digitized mammograms using morphology and topography-based computer aided detection schemes. Invest Radiol. 1998;33:746–751. doi: 10.1097/00004424-199810000-00006. [DOI] [PubMed] [Google Scholar]
- 2.Mascio LN, Hernandez J, Logan CM. Automated analysis for microcalcifications in high resolution digital mammograms. Proc SPIE. 1993;1898:472–479. doi: 10.1117/12.154533. [DOI] [Google Scholar]
- 3.Betal D, Roberts N, Whitehouse GH. Segmentation and numerical analysis of microcalcifications on mammograms using mathematical morphology. Br J Radiol. 1997;70:903–917. doi: 10.1259/bjr.70.837.9486066. [DOI] [PubMed] [Google Scholar]
- 4.Vincent L: Morphological grayscale reconstruction: Definition, efficient algorithm and applications in image analysis. Proceedings, Computer Vision and Pattern Recognition ’92, Champaign IL, June 1992, pp 633–635
- 5.Pal NR, Pal SK. Entropy thresholding. Signal Process. 1989;16:97–108. doi: 10.1016/0165-1684(89)90090-X. [DOI] [Google Scholar]