In this paper, we present a new method for simultaneously registering mammograms and detecting abnormalities. We assume that pixels can be divided into two classes: normal tissue and abnormalities (lesions). We define the registration constraints as a mixture of two distributions which describe statistically image gray-level variations for both pixel classes. The two distributions are weighted at each pixel by the probability of abnormality presence. Using the Maximum A Posteriori, we estimate the registration transformation and the probability map of abnormality presence at the same time. We illustrate the properties of our technique with some experiments and compare it with some classical methods.
展开▼
