VMRSimp

Massive Model Simplification

Ironically, most algorithms for simplifying polygonal models cannot handle the really big ones. Generally, this results from model accesses that are not coherent in memory. We have adapted our RSimp algorithm to build locality as it simplifies, allowing efficient simplification of models consisting of over 50 million faces in 32 bit architectures, and even larger models in 64 bit architectures. Because the simplification is globally adaptive, it remains quite accurate even at these large sizes.

Project members: Prasun Choudhury (NU ME), Donald Fong (UVA) & Ben Watson.

Sponsors: NSF Career award 0093172.

Thanks to: the Stanford 3D scanning repository, & the Digital Michaelangelo project.

Publications

P. Choudhury & B.A. Watson (2002). Completely adaptive simplification of massive meshes (pdf). Technical report NWU-CS-02-09.

V. Salamon, P. Lu, B.A. Watson, D. Brodsky & D. Gomboc (2001). A case study of improving memory locality in model simplification: metrics and performance (pdf). Proc. High Performance Computing (Hyderabad, India, December), 137-148. Springer Verlag, ISBN 3-540-43009-1.

Imagery

	Simplifications of Lucy Here are simplifications of the 28M face Lucy model to 1K, 10K and 100K face output sizes. All these simplifications were performed in roughly 10 mins on a 1GHz machine.
	Simplifications of David Here are simplifications of the 56M face David model to 1K, 10K, and 100K output sizes. All these simplifications were performed in roughly 20 mins on a 1GHz machine.