A multi - scale probabilistic network model for detection , synthesis and 13 compression in mammographic image analysis

20 We develop a probabilistic network model over image spaces and demonstrate its broad utility in mammographic image analysis, 21 particularly with respect to computer-aided diagnosis. The model employs a multi-scale pyramid decomposition to factor images across 22 scale and a network of tree-structured hidden variables to capture long-range spatial dependencies. This factoring makes the computation 23 of the density functions local and tractable. The result is a hierarchical mixture of conditional probabilities, similar to a hidden Markov 24 model on a tree. The model parameters are found with maximum likelihood estimation using the expectation-maximization algorithm. The 25 utility of the model is demonstrated for three applications: (1) detection of mammographic masses for computer-aided diagnosis; (2) 26 qualitative assessment of model structure through mammographic synthesis; and (3) compression of mammographic regions of interest. 27  2003 Elsevier Science B.V. All rights reserved.