Image Analysis

This final chapter gathers two approaches to the analysis of image-related datasets, the former pertaining to the area of so-called supervised classification where some datapoints are identified as belonging to a certain group (as opposed to the unsupervised classification of mixture data in Chapter 6) and the latter pertaining to pattern detection and image correction. Classification is operated via a probabilistic version of the k-nearest-neighbor method, while pattern detection is based on Potts modeling.

Image analysis has been a very active area for both Bayesian statistics and computational methods in the past thirty years, so we feel it well deserves a chapter of its own for its specific features. This is also the only introduction to spatial statistics we will provide in this book, and we thus very briefly mention Markov random fields, which are extensions of Markov chains to the spatial domain. A complete reference on this topic is Møller (2003).