Automatic Classification of Clustered Microcalcifications by a Multiple Classifier System

Mammography is a not invasive diagnostic technique widely used for early detection of breast cancer. One of the main indicants of cancer is the presence of microcalcifications, i.e. small calcium accumulations, often grouped into clusters. Automatic detection and recognition of malignant clusters of microcalcifications are very difficult because of the small size of the microcalcifications and of the poor quality of the mammographic images. Up to now, mainly two kinds of approaches have been proposed to tackle this problem: those performing the classification by looking at the features of single microcalcifications and those based on the classifications of clusters, which in turn use features characterizing the spatial distribution of the microcalcification in the breast. In this paper we propose a novel approach for recognizing malignant clusters, based on a Multiple Classifier System (MCS) which uses simultaneously the evidences obtainable from the classification of the single microcalcifications and from the classification of the cluster considered as a whole. The approach has been tested on a standard database of 40 mammographic images and revealed very effective with respect to the single experts.