Detection of microcalcifications in mammograms has received much attention from researchers and public health practitioners in these last years. The challenge is to quickly and accurately overcome the development of breast cancer which affects more and more women through the world. Microcalcifications appear in a mammogram as fine, granular clusters, which are often difficult to identify in a raw mammogram. Although, a variety of techniques have been proposed in the literature to enhance and automatically detect microcalcifications, but no method gives full satisfaction and clinically acceptable results. In this paper, we propose different wavelet based techniques for automatically microcalcifications detection. In a first time, we propose a pre-processing step to enhance the mammograms. In a second time, we propose different wavelet based techniques; from undecimated wavelet transform to multi-scale product, including the wavelet packets transform, the one-dimensional modulus maxima wavelet transform, and the two-dimensional to multi-scale product. Simulations are operated on Mini-Mammographic Image Analysis Society (MIAS) database and the results are presented and compared to some relative works. We have shown that the proposed approach is competitive with the best of the state of the art. The enhancement and the different wavelet based techniques proposed are the major contributions of this work.

Keywords: Breast cancer, mammography, wavelets, wavelet packets, modulus maxima, multi-scale product.

(P1150833305, 1.24 MB)