Biostatistics Seminar: BST 290: John Kornak (UC San Francisco)
DATE: Tuesday, December 2, 2014
LOCATION: 3106 Math Sciences Building (3rd floor, Math Dept)
SPEAKER: JOHN KORNAK, Associate Professor, Department of Epidemiology and Biostatistics, UC San Francisco
Title: “Bayesian image analysis in Fourier space, with applications in medical imaging”
Abstract: Image analysis is an extensive field that includes noise-reduction, de-blurring, feature enhancement, and object detection/identification, to name a few. Bayesian image analysis has played an increasing part in the field for at least 25 years, largely due to its structured approach to balancing a priori expectations of image characteristics with a model for the image degradation process (noise, blurring etc.). I will provide background on Bayesian image analysis, and in particular discuss the major role played by Markov random fields as prior distributions. I will subsequently describe my reformulation of the conventional Bayesian image analysis paradigm in Fourier space; spatially correlated processes, that are relatively difficult to model and compute in conventional image space, are more efficiently modeled as a set of independent processes across Fourier space. The Fourier space independence property leads to easy model specification and relatively fast and direct computation that is on the order of that for simple deterministic filtering methods. We will give specific examples of applications in medical imaging, and contrast Bayesian image analysis results in Fourier space with those in conventional image space.