Date of Award
Doctor of Philosophy (PhD)
Generating surfaces from spatial curves with topology constraints is a fundamental task in computer graphics. It has numerous graphics applications, such as computer-aided design (CAD), product prototype and biology shape analysis. Though classic and widely-needed, topology-aware surface reconstruction from multiple spatial curves stays a daunting task in computer graphics because of the enormous ambiguity in the geometry and topology of a surface. In this dissertation, we propose three methods tailored for reconstructing surfaces with specific topology properties. The first method obtains a manifold triangulation of multiple, arbitrarily oriented 3D polygons. The surface is guaranteed to be manifold and optimizes a user customizable metric. The second and third method study the reconstruction problem on curves from cross-sections. In particular, we offer topological control during the reconstruction. A family of topologies is explored and scored during this process. The second work focuses on a novel complete framework for reconstructing surfaces with user specified topology and the third work extends the topology exploration of this framework from handling two materials to multiple.
Tao Ju, Nathan Carr, Yixin Chen, Yasutaka Furukawa,
Available for download on Friday, May 15, 2116