基于一种可变分类阈值的SPIHT算法

针对传统小波变换过程复杂和多级树集合分裂算法（Set Partitioning In Hierarchical Trees,SPIHT）编码过程重复运算、存储量大的问题,提出了一种新的可变阈值的SPIHT算法。该算法利用可变阈值对SPIHT算法中不重要的像素集合链表（List of Insignificant Pixels,LIP）、系数列表（List of Insignificant Sets,LIS）进行分类,得到一种更有效的编码算法。仿真结果表明,该算法针对不同图像的大小、分解级数都能得到很好的峰值信噪比（Peak Signal-to-Noise Ratio,PSNR）。

A New SPIHT Algorithm Based on Variable Sorting Thresholds

Aiming at the problems of complicated convolution process of wavelet transform,repeated calculations and a great of needed memories of SPIHT algorithm,we present a more efficient implementa-tion of the SPIHT algorithm by using variable thresholds to sort the list of insignificant pixels(LIP)and the list of the insignificant sets(LIS).Simulation results show that this algorithm can work well on different image sizes and different decomposition levels.