一种基于预测分类的小波图像编码

Wavelet coding based on predicted classification.

对于小波图像编码,如何有效地组织小波域系数是提高图像压缩效果的关键.本文在研究小波域系数统计特性的基础上,提出一种用于图像压缩的预测分类模式.与基于零树结构的编码不同的是它充分利用了小波域中"重要"系数的各种相关性,并通过结合提出的种子膨胀算法实现系数的分类输出.实验结果表明该方法所取得的压缩效果要优于基于零树结构的图像编码,同时又具有较强的鲁棒性.

Organizing the wavelet coefficients effectively is pivotal to the performance of image coding based on wavelet transform. With studying of the statistical properties of wavelet coefficients, a predicted classification model for image coding is presented. Contrary to the Zerotree based method, this model takes the advantage of dependencies between significant coefficients in the wavelet domain. By using this model combined with the proposed seeddilation algorithm, our experiment shows that the results of image coding have better reconstruction quality and high robustness. 