Title :
Breast mass segmentation on digital mammograms by a combined deterministic annealing method
Author :
Cao, A.Z. ; Song, Q. ; Yang, X.L. ; Liu, S.
Author_Institution :
Sch. of Electr. & Electron. Eng., Nanyang Technol. Univ., Singapore
Abstract :
Segmentation plays a vital role in digital mammographic mass detection. In this paper, we developed one breast mass segmentation scheme which is based on the concept of deterministic annealing (DA) and incorporated with a gain adaptive contrast enhancement (GACE) preprocessing technique. The probability distribution of gray levels within the region of interest(ROI) after GACE filter are estimated and a series of image partitions are obtained according to the intensity proximity as the temperature decreases. Compared with the segmentation results of DA without enhancement, and DA with histogram enhancement, the experimental results show that the combined DA and GACE segmentation scheme works more efficiently.
Keywords :
cancer; image enhancement; image segmentation; mammography; medical image processing; statistical distributions; tumours; ROI; breast mass segmentation; deterministic annealing method; digital mammogram; digital mammographic mass detection; gain adaptive contrast enhancement preprocessing technique; gray level; histogram enhancement; image partition; intensity proximity; probability distribution; region of interest; Annealing; Breast cancer; Breast tissue; Clustering algorithms; Filters; Histograms; Image databases; Image segmentation; Probability distribution; Temperature distribution;
Conference_Titel :
Biomedical Imaging: Nano to Macro, 2004. IEEE International Symposium on
Print_ISBN :
0-7803-8388-5
DOI :
10.1109/ISBI.2004.1398785