Physics-based inverse problems in medical image analysis
Friday, September 23, 10:30AM – 11AM
George Biros, ICES
Medical image analysis refers to computer-assisted interpretation of clinical images. For example, images acquired using ultrasound, computer assisted tomography, or nuclear magnetic resonance. This interpretation includes but goes beyond standard image and signal analysis (e.g., denoising, compression, reconstruction) and it mainly refers to segmentation, registration, classification (diagnosis) and prediction (prognosis).
Similar to weather prediction in which implicit (radar) or explicit (stations) measurements of various field quantities (velocity, pressure) are combined with sophisticated mathematical models encapsulating conservation laws, physics-based inverse problems in medical image analysis is an attempt to provide a mathematical and algorithmic framework for computer-aided assisted diagnosis and prognosis.
Physics-based inverse problems are image processing techniques that use prior knowledge related to the physiology (mechanics, biochemistry, electrophysiology) of the imaged tissue.
The related mathematical framework is theory and algorithms for optimization of systems governed by partial differential equations. In this talk, I will give details on the formulation for two problems: cardiac motion estimation and deformable registration of images with brain tumors with a normal brain atlas.
The talk will be at 10:30 am in the ICES 6th floor seminar room (POB 6.304). Coffee and cookies will be provided. We hope to see you there.
Hosted by Ivo Babuska