ABSTRACT
This chapter focuses on problems that can be posed as labelling or clustering problems in which the solution is a set of labels assigned to image pixels or features. It presents how Hidden Markov random field (HMRF) models generalize standard mixture models and proposes an inference procedure using variational approximation and illustrates the framework with two real medical image applications. The chapter deals with image data where each pixel is associated with a univariate observation. Image analysis includes a variety of tasks such as image restoration, segmentation, registration, visual tracking, retrieval, texture modelling, classification and sensor fusion. For image analysis or spatial data clustering, dependencies or contextual information can be taken into account using HMRF models, which can be seen as a generalization of standard mixture models. Many applications are related to image analysis, but other examples include population genetics and bioinformatics.
