The development of cardiac image analysis and modeling tools is
becoming very important due to the
ever increasing cardiac imaging necessary for the timely diagnosis of
heart diesease. In this talk we will present the steps and methods
necessary for developing a clinically useful image analysis and modeling
software for the analysis of Tagged MRI images.
We will begin with low level image processing based on Gabor Filters for
the extraction of Tagging lines, followed by deformable models and
learning methods for the extraction of the heart wall boundaries, the 3D
reconstruction of the cardiac geometry, and the motion analysis of the
heart. Then, we will present finite element methods and stochastic
methods for the estimaiton of the
stress-strain realtionship in the heart and the estimation of fiber
orientation. Finally, we will discuss methods for modeling blood flow.
We will conclude by giving the spectrum of future problems and
challenges that need to be addressed in these areas.
Audio (MP3 File, Podcast Ready)
Copyright. All Rights Reserved.