3D mesh compression

In this chapter, we discuss 3D compression techniques for reducing the delays in transmitting triangle meshes over the Internet. We first explain how vertex coordinates, which represent surface samples may be compressed through quantization, prediction, and entropy coding. We then describe how the connectivity, which specifies how the surface interpolates these samples, may be compressed by compactly encoding the parameters of a connectivity-graph construction process and by transmitting the vertices in the order in which they are encountered by this process. The storage of triangle meshes compressed with these techniques is usually reduced to about a byte per triangle. When the exact geometry and connectivity of the mesh are not essential, the triangulated surface may be simplified or retiled. Although simplification techniques and the progressive transmission of refinements may be used as a compression tool, we focus on recently proposed retiling techniques designed specifically to improve 3D compression. They are often able to reduce the total storage, which combines coordinates and connectivity, to half-a-bit per triangle without exceeding a mean square error of 1/10,000 of the diagonal of a box that contains the solid.