Agenda

Abstract

Learning Models of Structure
Kobus Barnard, Ph.D.

The form of objects, organisms, and organs is often closely linked to their function. Going beyond simple recognition, being able to capture and quantify three-dimensional structure has many important applications. Examples include recognizing the pose (and thus actions) of a person, inferring the species of an organism, and classifying the phenotype of cells with different genotypes grown under different conditions.

In this talk I will outline a general approach to modeling structure, and simultaneously inferring it and imaging system parameters using Bayesian methodology. The inference approach applies equally well to fitting an existing model to an individual, as well as to learning the model itself through estimating meta-parameters for a population of examples from a class. I will illustrate the approach in diverse applications that include a stochastic L-system model for microscopic fungus from the genus Alternaria, a simple model for common tables, and a proposed model for brain neurons grown in culture.

 
Return to Agenda.