Detection of areas containing microcalcifications in digital mammograms
Mario Muštra, Mislav Grgić · International Conference on Systems, Signals and Image Processing · 2014
Microcalcifications are tiny calcium deposits present in the breast tissue. They are proved to be an important early sign of a process which started to happen and could be a development of a breast cancer. Because of that computer-aided detection systems (CADe) for detection of microcalcifications with low false positive rate can be very useful and helpful for breast cancer prediction. Accurate detection of microcalcifications is a very challenging task because of their visual property which is very similar to the noise in the image. To be able to perform accurate detection it is necessary to remove the background influence and at the same time preserve microcalcifications in the image. In this paper we propose a method for detection of presence of microcalcifications using the contrast enhancement of microcalcifications against the background tissue. For background suppression we use combination of wavelet filtering and grayscale morphology. The results of detection of suspicious areas which could contain microcalcifications are shown to be good in comparison to other methods which do not use manual selection of suspicious areas.