一种改进的SPIHT算法

A modified SPIHT algorithm

利用特定的滤波器对自然风景图像、人物图像进行小波分解后 ,在低频区域中 ,小波系数的符号绝大部分为正。利用这一特征 ,对多级树集合划分 (SPIHT)算法进行了改进 ,使得在同样的压缩比下 。

Set partitioning in Hierarchical Trees(SPIHT)coding,introduced by A.Said and W.A.Pearlman is a fast and efficient technique for image compression. Most signs of the coefficients in the highest pyramid level are positive after wave let transformation. If the 9/7-tap filter of [3] is selected,the ratios of posi ti ve coefficients to the total coefficients in the highest pyramid level are very high.Thus the performance by exploiting this character can be improved.