摘要： PVDS算法因搜索固定数量的纹波导致搜索范围过大，编码效率较低。针对该问题，提出一种基于Voronoi图的自适应纹波搜索算法APVDS。通过实验确定一组合理的阈值，每搜索一个纹波就根据阈值判断是否达到搜索停止条件，由此减少所需搜索的纹波数。仿真实验结果表明，自适应搜索到2个纹波后，APVDS与PVDS算法的编码质量基本相同，但平均搜索范围明显缩小，平均编码时间也相应减少。

关键词: 矢量量化, 快速码字搜索, 自适应搜索, PVDS算法, 主成分分析

Abstract: Planar Voronoi Diagram Search(PVDS) algorithm always searches a fixed number of ripple waves, so its search range is too large. This paper proposes an adaptive search algorithm for ripple waves based on Voronoi diagram named Adaptive PVDS(APVDS). A set of reasonable thresholds are found based on experiments. Whether to terminate the whole search flow after finishing each ripple search depends on the thresholds. Experimental results show that APVDS achieves almost the same Peak Signal to Noise Ratio(PSNR) compared with PVDS, and reduces the search space and the coding time.

Key words: Vector Quantization(VQ), fast codeword search, adaptive search, Planar Voronoi Diagram Search(PVDS) algorithm, Principal Component Analysis(PCA)

中图分类号: 

• TN919.8