Breast Tissue Classification in Mammograms Using ICA Mixture Models

In this paper we present a novel method for recognizing all kinds of abnormalities in digital mammograms using Independent Component Analysis mixture models and two sets of statistical features based on texture analysis. Our approach is concentrated on finding the ICA mixture model parameters that describe in an exclusive and effective way the abnormal and the normal tissue, and with the aid of a supervised probabilistic classifier we are able to successfully recognize suspicious regions in mammograms. Extensive experiments using the MIAS database have shown great accuracy of 98.33% in classifying an unknown regions of suspicion as abnormal and 62.71% as healthy tissue.