一种基于相关性预测的小波图像编码

Wavelet image coding based on dependency predication

一幅图像经小波变换后,不同子带的变换系数之间有着一定的相关性.为了更好地利用这个特点,提出了一种新的图像压缩算法.该方法通过对高频子带的系数变换重排,将每个子带内的重要系数尽可能地集中到相应子带的特定区域.然后采用子带间重要系数预测的方法,对小波变化系数进行编码.实验结果表明,通过子带内的系数重排,在几乎不影响压缩时间的情况下,较大地提高了图像的压缩率.

There are certain correlations between wavelet coefficients in specific subbands after wavelet transform of an image. This paper proposes a new image compression method wherein coefficients in highfrequency subbands are reordered by means of transform. The significant coefficients are then concentrated in the specific area of the corresponding subband. After this, a predication of significant coefficients is employed to compress those coefficients. The experiment results showed that after coefficients reordering, entailing only little increase in processing time, the compression rate was improved remarkably.