Automatic segmentation and detection of mass in digital mammograms
Nor Ashidi Mat Isa, Ting Shyue Siong · International Conference on Signal Processing · 2012
This paper presents an automated system for mass segmentation and detection in mammograms. Initially, breast segmentation is applied to separate the breast and non-breast area. Then, image enhancement is employed to improve the contrast of the tissues structure in mammograms. Finally, constraint region growing based on local statistical texture analysis is applied to detect and segment out the mass from the mammograms. The system is develop and evaluated with 322 mammograms from Mammographic Image Analysis Society Database. The verification results show that the proposed technique has a sensitivity of 94.59% and the number of false positive per image is 3.90.