Quality of Microcalcification Segmentation in Mammograms by Clustering Algorithms
Breast cancer remains a leading cause of death among women worldwide. Mammography is one of the non-invasive methods to find breast tumors, which is very useful in the detection of cancer. Microcalcifications are one of the anomalies of this disease, and these appear as small white spots on the images. Several computer-aided systems (CAD) have been developed for the detection of anomalies related to the disease. However, one of the critical parts is the segmentation process, as the rate of detection of anomalies in the breast by mammography largely depends on this process. In addition, a low detection endangers women’s lives, while a high detection of suspicious elements have excessive cost. Hence, in this work we do a comparative study of segmentation algorithms, specifically three of them derived from the family of c-Means, and we use the NU (Non-Uniformity) measure as a quality indicator of segmentation results. For the study we use 10 images of the MIAS database, and the algorithms are applied to the regions of interest (ROI). Results are interesting, the novel method of sub-segmentation allows continuous and gradual adjustment, which is better adapted to the regions of micro calcification, and this results in smaller NU values. The NU measure can be used as an indication of quality, which depends on the number of pixels and the homogeneity of the segmented regions, although it should be put in the context of the application to avoid making misinterpretations.
Unable to display preview. Download preview PDF.
- 4.Hernandez-Cisnero, R.R., Terashima-Marn, H.: Evolutionary neural networks applied to the classification of microcalcification clusters in digital mammograms. Proc. IEEE Congr. Evol. Comput., 2459–2466 (2006)Google Scholar
- 7.Gonzalez, R.C., Woods, R.E.: Digital Image Processing, 2nd edn. Publishing House of Electronics Industry, Beijing (2007)Google Scholar
- 9.MacQueen, J.B.: Some methods for classification and analysis of multivariate observations. In: Proc. 5th Berkeley Symposium on Mathematical Statistics and Probability, pp. 281–297. University of California Press, Berkeley (1967)Google Scholar
- 10.Bezdek, J.C., Keller, J., Krishnapuram, R., Pal, N.R.: Fuzzy models and algorithms for pattern recognition and image processing, Boston, London (1999)Google Scholar
- 11.Jain, A.K., Dubes, R.C.: Algorithms for Clustering Data. Prentice-Hall, Englewood Cliffs (1998)Google Scholar
- 17.Serra, J.: Images analysis and mathematical morphological. Academic Press, New York (1982)Google Scholar
- 20.Ojeda-Magaña, B., Quintanilla-Domínguez, J., Ruelas, R., Andina, D.: Images sub-segmentation with the PFCM clustering algorithm. In: Proc. 7th IEEE Int. Conf. Industrial Informatics, pp. 499–503 (2009)Google Scholar