RUI: Geometric Techniques for Quadrilateral and Hexahedral Mesh Generation with Applications in Medical Imaging

Rutgers University New Brunswick, New Brunswick NJ

Investigators

Abstract

ABSTRACT 0204293 Suneeta Ramaswami Rugers U New Brunswick The broad goal of this project is the complete development of robust algorithms for the modeling and visualization of data in a specific application domain, namely medical imaging. The focus of the proposal is on the generation of quadrilateral and hexahedral meshes for medical data obtained from structural studies (MR and CT) of human organs. The generation of good quadrilateral and hexahedral (quad/hex) meshes is not well-understood and several important questions remain open. By exploiting the geometry that is central to these problems, we aim to develop algorithms for generating guaranteed-quality surface and volume meshes composed of quadrilateral and hexahedral elements. Algorithmic techniques and data struc- tures from computational geometry will be utilized towards this end. We also investigate adaptive quad/hex meshes, i.e., meshes designed to have varying levels of refinement depending on factors such as the desired level of detail in a localized part of the input domain. The algorithms will be designed for and tested on medical imaging data. Improved surface and volume meshing algorithms will directly impact automated clinical analysis of medical data, which is typically carried out by procedures that require finite element simulations on discrete anatomical models. Some examples of such procedures are elastic matching and viscous uid ow. Our work will be done in collabo- ration with researchers in the area of brain image analysis in the Department of Radiology at the University ofPennsylvania.