基于内容和系数划分的嵌入式零数编码算法

Study on embedded zero-tree coding algorithm based on combination of context and coefficient partition

提出了一种采用系数块划分、零树和上下文模型的嵌入式图像编码方法.它不同于EZW等算法对有效小波系数的有效性独立编码的思想,而是将相邻且具有相同有效性的系数结合成一个数据块一起编码.在时间、空间的层数和熵编码相同条件下,新的算法与EZW相比,在相同的PSNR下,压缩效率提高大约10%.

In this paper, an embedded coding algorithm based on coefficient-block partition, zero-tree and context is proposed, unlike traditional algorithms which often independently coding about the wavelet significant coefficients, such as EZW etc, the algorithm combines the adjacent coefficients having the same validity as a coefficient block, then code together. Compares with EZW, the compression ratio of the new algorithm can be improved about 10%, same under the same temporal and spatial scales, the same entropy coding method and the PSNR.