基于HVS的DWT图像压缩混合编码算法

为提高数字图像在高压缩比下重构图像的质量，分析研究现有离散小波变换（discrete wavelet transform，DWT）图像压缩的相关算法以及小波系数特性，提出一种基于人眼视觉系统（human visual system，HVS）特性对小波系数进行加权，结合小波各子带系数特性，采用差分脉冲编码调制（differential pulse code modulation，DPCM）与多级树集合分裂排序（set partitioning in hierarchical trees，SPIHT）相结合的编码方法．经视觉信息保真度分析实验验证，与传统的嵌入式零树小波（embeddedzero—tree wavelet，EZW）和SPIHT算法相比，在相同的比特率下该算法重构的图像具有更好的主观视觉效果．

DWT Image Compression Hybrid Coding Algorithm Based on HVS

In order to improve the reconstructed image quality of digital images at high compression ratio, the existing discrete wavelet transform （DWT） algorithms for image compression and wavelet coefficient features were studied. Encoding method was based on human visual system （HVS） characteristics of the wavelet coefficients weighted, combined with wavelet coefficient characteristics of each sub-band, using differential pulse code modulation （DPCM） and set partitioning in hierarchical trees （SPIHT）. The visu- al information fidelity （VIF） analysis experiment showed that the reconstructed image of the proposed al- gorithm had better subjective visual effect, compared with the traditional embedded zero-tree wavelet （EZW） and the SPIHT algorithm at the same bit rate.