基于新型嵌入小波编码策略的图像压缩算法研究

在一系列新量化编码算法中，由J．M．shapiro提出的嵌入零树小波编码(简称为EZW)被认为是最先进的方法之一，然而，它也存在明显不足之处：为了确定小波系数是零树根还是孤立零，必须利用小波系数间存在的双亲-子女(Parent-Children)关系去搜索扫描其所有后代，而这必然会导致扫描高频子带所产生的二进制符号流中存在大量冗余，从而大大影响其编码效率。提出了一种基于新型嵌入小波编码策略的图像压缩算法，该算法能够通过以下措施提高EZW编码方案工作效率：①采用全新的嵌入小波编码策略，不再利用导致计算量剧增的双亲-子女(Parent-Children)关系；②进行嵌入小波编码时，不再考虑ZTR(零树根)和IZ(孤立零)符号。实验结果表明：提出的新图像编码算法在压缩／解压缩速度、图像复原质量等方面均明显优于EZW编码方案。

A New Embedded Wavelet Coding Strategy for Image Compression

An image compression method based on new embedded wavelet coding strategy is presented. The well-known Embedded Zerotree Wavelet(EZW)coder uses Successive Approximation Quantization(SAQ)process and zerotree structure of wavelet coefficients, it scans iteratively whole wavelet coefficients during the SAQ process by using the parent-children relation of the coefficients to multiresolution image decomposition, which decreases the coding efficiency considerably. In our proposed method, the significant coefficients whose absolute values are greater than a threshold are coded by using whole coefficients, but the parent-children relation is not used and ZTR and IZ symbols are not considered for coding. Therefore the time-consuming problem can be avoided. Experimental results show that our proposed coding method performs better than that of EZW on the aspects of PSNR, encoding time and decoding time.