Dissimilarity Application for Medical Imaging Classification

This work concerns the development of an automatic classification system which could be useful for radiologists in the investigation of cancer from breast diagnostic images. The software has been designed in the framework of the MAGIC-5 collaboration. In the traditional way of learning from examples of objects the classifiers are built in a feature space. However, an alternative way can be found by constructing decision rules on dissimilarity (distance) representations. In this recognition process a new object is described by its distances to the training samples. The use of dissimilarities is especially of interest when features have an insufficient discriminative power. In the automatic classification system the suspicious regions with high probability to include a lesion are extracted from the image as regions of interest (ROIs). Each ROI is characterized by some features extracted from co-occurrence matrix containing spatial statistics information on ROI pixel grey tones. A dissimilarity representation of these features is made before the classification. A feed-forward artificial neural network is employed to distinguish pathological records, from nonpathological ones by the new features. The results obtained in terms of sensitivity and specificity are presented.