北京工业大学计算机学院,北京,100022
纸质出版:2004
移动端阅览
刘 波, 张鸿宾. 三角形网格的分块DCT压缩[J]. 电子学报, 2004,32(2):181-185.
LIU Bo, ZHANG Hong-bin. Block DCT Compression of Triangle Meshes[J]. Acta Electronica Sinica, 2004, 32(2): 181-185.
在现有的代表性三角形网格压缩方法中
先采用一定的网格遍历方法来压缩连接信息
同时用遍历路径上的相邻顶点来对每个顶点的几何坐标进行预测
以压缩几何信息.其主要缺点是只利用了遍历路径上的相邻顶点来进行预测
并没有充分去掉顶点间的相关性.其实在空间中一定局部范围内
所有顶点的坐标间都存在着一定的相关性
这些顶点虽然在空间上相邻
但并不一定在遍历路径上相邻.和图像压缩标准JPEG的思路类似
本文提出一种新的基于分块DCT的网格几何信息压缩方法.先将网格划分成很多基本同样大小的块
利用每个块内的所有顶点按遍历次序排列成一维序列后
坐标呈周期性分布的事实
采用一维DCT变换来去除块内顶点间的相关性.实验表明
分块DCT方法取得了较好的几何信息压缩性能.
In some representative methods of triangle mesh compression
the connectivity is encoded using some kinds of traversal methods
the geometry is encoded by predicting the coordinates of each vertex with neighboring vertices along the traversal path.Their main drawback is the prediction only reduces the correlation between each vertex and its neighboring vertices in the traversal path.The coordinates of all vertices in a local spatial region have strong correlation.although these vertices are neighboring in space
they are usually not neighboring in the traversal path.Inspired by image compression standard JPEG
a novel block DCT geometry compression method is proposed.The input mesh is partitioned into many blocks.Given the fact that if all vertices in a block are ordered in traversal order
their coordinates distribute periodically
we use 1-dimentional DCT transform to reduce the correlation among all vertices in a block.The experimental results show the block DCT method has good geometry information compression performance.
0
浏览量
1094
下载量
0
CSCD
关联资源
相关文章
相关作者
相关机构
京公网安备11010802024621