Share Email Print
cover

Proceedings Paper

Piecewise compression of large mesh
Author(s): Aihong Qin; Hua Xiong; Jiaoying Shi; Zhen Liu
Format Member Price Non-Member Price
PDF $14.40 $18.00

Paper Abstract

Large and detailed 3D polygon mesh with standard representation results in files of gigantic size. The need for more compact representations and a parallel implementation is clear. But the compressed gigantic mesh applied in parallel rendering to achieve high performance is still an unexplored area. In this work, we present a mesh compression scheme employed in parallel rendering system. It includes two parts: Mesh segmentation and segments compression. Firstly, the multilevel graph partitioning idea is adopted to separate the mesh into large patches with less curvature. Then the large patches are farther partitioned with MeTiS[1] to get small patches with a balanced vertex counts. As the multi-level mesh produced by the feature preserved mesh simplification procedure, our segmentation algorithm takes both the segment's flatness and balanced vertex counts into account. Secondly, each patch is compressed separately. The vertices in the patch are classified into two distinguish types, namely boundary vertex and inner vertex, different compression algorithm are applied to them. The boundary vertexes are compressed with a novel compression algorithm PMC proposed in this work. To avoid decoding the whole boundary during every frame rendering, the boundary vertexes are compressed piecewise. The successive boundary edges shared by the same patches can act as a piece element when it is compressed. During sorting only the encoded boundary in the view-frustum needs to be loaded and decompressed. Experiments show that the encoded mesh can be partitioned in compression-domain in parallel rendering system. It reduces the communication bandwidth requirement significantly.

Paper Details

Date Published: 18 January 2006
PDF: 10 pages
Proc. SPIE 6066, Vision Geometry XIV, 60660P (18 January 2006); doi: 10.1117/12.642727
Show Author Affiliations
Aihong Qin, Zhejiang Univ. (China)
Shanxi Univ. (China)
Hua Xiong, Zhejiang Univ. (China)
Jiaoying Shi, Zhejiang Univ. (China)
Zhen Liu, Zhejiang Univ. (China)


Published in SPIE Proceedings Vol. 6066:
Vision Geometry XIV
Longin Jan Latecki; David M. Mount; Angela Y. Wu, Editor(s)

© SPIE. Terms of Use
Back to Top