摘要： 针对小波变换在压缩高维空间数据特征上存在的明显不足，结合多尺度几何分析中的wedgelet理论，提出一种新的三维体数据压缩算法。将wedgelet理论扩展到三维，利用其多尺度多分辨率分析能力及体数据的几何正则性对三维体数据进行压缩，较好地保留了体数据的轮廓曲面特征。通过提升小波及DCT方法对3种不同体数据的压缩实验显示，该算法压缩率较高、重构效果比较理想。

关键词: wedgelet理论, 数据压缩, 八叉树, 小波

Abstract: Aiming at wavelet based algorithms having significant shortcomings in their treatment of high dimension data characteristic, this paper presents a novel 3D volume data compression algorithm based on extend wedgelet of multiscal geometry analysis. The algorithm uses the excellent multiscal multiresolution analysis performance of extend wedgelet and the geometrical regularity of volume data to compress the volume data, and as a result, the volume data contour curve surface characteristics are saved effectively. Experimental results show that the algorithm outperforms adaptive wavelet-based algorithm and DCT algorithm, and is effective for compressing the volume data.

Key words: wedgelet, data compression, octree, wavelet

中图分类号: 

• TP391