Chair: Jason Lawrence, Todd Zickler, Kevin Skadron and William Walker
Advisor:
OLSSON 236D, 13:00:00
A Ph.D. Proposal
ABSTRACT
While many techniques for measuring the 3D geometry of physical objects have been proposed, the problem of jointly measuring shape and appearance has not been fully addressed. The few systems designed for measuring both shape and appearance rely on error-prone techniques for aligning 2D
images with existing 3D models and place strong assumptions on the geometry and reflectance of
target objects. I propose to investigate practical methods for the synchronous capture of 3D shape
and surface appearance. The major contributions of this work will include (1) new optical devices
that limit calibration effort and improve data quality, (2) new algorithms for efficient image-based 3D reconstruction, and (3) a prototype system for demonstrating and evaluating my
approach.