Multi-scale Geometry Constrained
Shape Diffeomorphism
Guangyu Zou
Dept. of Computer Science
Wayne State University
Tuesday,January20, 2009
3:00pm Rm 110 Purdy-KresgeLibrary
Abstract:
With the advent of active and passive 3D sensing techniques, 3D geometric data now plays vital roles in many applications. In order to accurately represent the geometry of 3D objects, extracting a set of descriptive geometric features becomes crucial. Inspired by the observation that semantically meaningful geometric features can only be captured at inherent spatial scales, we embed 3D shapes in a scale space for further analysis. In this talk, I will first exploit this hidden dimension of 3D geometry, i.e., the geometric scale variability, for a stable, scalable, compact shape representation which effectively supports matching and comparison of large-scale geometric data. Based on this multi-scale geometry representation, automatic non-rigid registration of 3D data is then solved for the essential role it plays in a wide range of applications and its technical challenges. In particular, shape deformation is studied in a more abstract space of diffeomorphisms, giving rise to a more general and in-depth understanding of 3D shape evolution. I will also demonstrate some applications of this framework to the multimodal neuro-imaging data analysis at the end.
Biography:
Guangyu Zou received his Ph.D. in Computer Science at WayneStateUniversity in 2008, where he was also a research assistant in the Graphics and Imaging Laboratory. Since 2009, he has been working as a postdoctoral research fellow in the Department of Computer Science at the State University of New York at Stony Brook. His research centers on the representation, processing and visualization of geometric data. Areas of interest include computer graphics, shape modeling, geometry processing, visualization, computer vision and medical imaging.
5143 Cass Avenue 431 State Hall Detroit, Michigan 48202 +1.313.577.2477 Fax +1.313.577.6868