基于上下文算术编码的非三角网格拓扑压缩

Connectivity compression for non-triangular meshes by context-based arithmetic coding

网格拓扑压缩方法是计算机图形学的基础算法。该文方法是单分辨率,主要针对非三角网格模型的拓扑信息作无损压缩。算法首先遍历网格的所有多边形得到操作系列;然后对操作系列作霍夫曼编码;再对霍夫曼编码结果作基于上下文长度可变的算术编码得到最后的压缩结果。相比于对非三角网格拓扑信息作压缩的压缩比很高的算法,该算法得到的压缩结果更好。此算法的另一个突出优点是在解码时间和空间上有了改进——新算法可以在接收一个多边形的编码后立即完成解码并抛弃这个编码,从而使得该算法特别适用于在线传输和解码的实时与交互应用场合。此外,该算法还可以处理有空洞和柄(handle)的模型。

The method that encodes the connectivity information for general polygon meshes is the foundation of graphics field.The algorithm in this paper is a single-resolution lossless compression method for mesh models,mainly for non-triangu- lar mesh models.By the method, all faces are encoded firstly to obtain operator series, then Huffman coder is applied to en- code these operator codes, and finally a context-based arithmetic coder is employed to encode the Huffman codes.This meth- od can provide a higher compression ratio for non-triangular meshes in comparison with the excellent algorithms previously proposed.The new method can save much of the decoding time and space,by introducing a decoding scheme so that the op- erator code can be immediately discarded as soon as it is decoded.Therefore, the decoding method can be well applied to the applications with online transmission and decoding.The algorithm is also capable of handling the meshes with holes and handles.