VIVARIUM Seminar: 3D Geometry Reconstruction from Image Data and Biomedical Applications

July 31, 2008
Location: T9204 West
Presenter: Hans-Christian Hege, Head of Visualization and Data Analysis Department, Zuse Institute Berlin (ZIB), Germany


A major source of input data in many biomedical applications are 2D, 3D and 4D image data. For quantitative analysis often a geometric reconstruction of the depicted anatomical entities is necessary. Often also meshes suitable for finite element computations are necessary. In this talk a pipeline will be presented that allows to reconstruct anatomies in a variety of applications and image data. The steps to be performed are image preprocessing, image registration, image segmentation, geometry reconstruction and grid generation.
The most difficult step is image segmentation. Numerous image segmentation methods have been invented. A key for automated techniques is utilization of pior information. In recent times statistical shape models have been successfully used for that purpose. Creation of statistical shape models and utilization in image segmentation therefore will be explained.
Finally various biomedical applications will be shortly sketched where these techniques have been successfully applied. This should give you an idea how valuable tools for biomedical sciences can be created by deliberate combination of techniques from image processing, image analysis, geometry processing, computer graphics and data visualization.

See presentation slides here.


Site by Faculty of Applied Science web team, Simon Fraser University