摘要： 为了提高小波嵌入压缩算法的速度，提出了可变深度嵌入小波压缩算法。由图像小波变换能量聚集和子带间的幅度衰减性等特性，在小波分层嵌入编码过程中，对不同的重要图，部分小波子带的系数幅度不会超过当前的阈值，从而可以忽略不计，即小波树的深度随着重要图的不同而改变，也就是变深度嵌入编码。实验结果显示，对于自然图像来说，不需要考虑的层中的小波系数幅值几乎全部小于当前的阈值，从而和EZW算法相比，编码不会影响压缩率（或图像质量），而编解码过程中的运算量明显下降。

关键词: 图像编码, 小波, 嵌入压缩, 变深度嵌入

Abstract: Image compressing by wavelet method is usually the embedded compression. Every wavelet coefficient is scanned many times in the compressing. It can spend more time than no embedded algorithm. The variable depth embedded compression of natural images is proposed in this paper. By the decay of the magnitude of the wavelet coefficients across subbands, the embedded algorithm can only scan parts of wavelet subbands, that is variable depth embedding. The experimental results show that the speed is faster than the previous schemes.

Key words: Images coding, Wavelet, Embedded compression, Variable depth embedding

中图分类号: 

• TN919.8