Detection of Microcalcification in Digital Mammograms Using Multi-Scale Products and Active Contour Model

Mammography is the most efficient method for breast cancer early detection. Clusters of microcalcifications are the early sign of breast cancer and their detection is the key to improve prognosis of breast cancer. Microcalcifications appear in mammogram image as tiny localized granular points, which is often difficult to detect by naked eye because of their small size.  Automatic and accurately detection of microcalcifications has received much more attention from radiologists and physician. An efficient method for automatic detection of clustered microcalcifications in digitized mammograms is the use of Computer Aided Diagnosis (CAD) systems. This paper presents a two dimensional wavelet-based multiscale products scheme for microcalcification detection in mammogram images. Initially, Mammogram image was decomposed by 2D wavelet transform into different frequency sub-bands, the low-frequency subband was suppressed and significant high frequencies features were reconstructed. The significant high frequencies features were obtained by multiscale products. An Active contour model was applied on reconstructed image and microcalcification nodules were segmented from resulting image. Preliminary results indicate that the proposed scheme is better in suppressing the background and detecting the microcalcification clusters.