Shape Representation in Practice

University Of California-Davis, Davis CA

Investigators

Abstract

ABSTRACT Proposal #0098169 Annamaria B Amenta University of Texas The award comprises two projects concerning shape reconstruction from real-world data. The first is to work on automating the process of finding a conformation of a molecular model to match an electron density map produced by X-ray diffraction. This is a important problem in protein biochemistry, currently solved by laborious user interaction. Our recent research on the construction of the medial axis should be useful for building a concise encoding of the shape of the density map, which we can search for features to match to the molecular model. The second project is to adapt a two-dimensional image processing technique to produce a high-resolution three-dimensional model of an object by combining many low-resolution point sets cap-tured by a laser range scanner. This would allow the automatic construction of very high-resolution models of real objects, for applications such as customized medical appliances. The key to adapting the technology will be developing a computational model of the scanner itself, a very non-traditional imaging device.