B0613
Title: Bayesian image analysis in Fourier space models and some relationships with Markov random fields
Authors: John Kornak - University of California, San Francisco (United States) [presenting]
Karl Young - University of California San Francisco (United States)
Eric Friedman - International Computer Science Institute Berkeley (United States)
Abstract: Bayesian image analysis can improve image quality by balancing a priori expectations of image characteristics with a model for the noise process. The conventional Bayesian model in the image space approach implements priors that describe inter-dependence between spatial locations. It can therefore be difficult to model and compute. However, similar models can be developed more conveniently as a large set of independent processes when considered in Fourier space. The originally complex high-dimensional estimation problem in image space is thereby broken down into a series of (trivially parallelizable) independent one-dimensional problems in Fourier space. Example implementations of a range Bayesian image analysis in Fourier space models will be shown. How these Fourier space models relate to Markov random field-based models that are commonly used in conventional Bayesian image analysis will also be discussed.
