基于DWT和视觉加权的图像质量评价方法研究

摘要

A new objective method WVWPSNR (wavelet and vision_weighted peak signal noise ratio) was presented. In which the method was based on DWT, and the sensitivity of HVS (human visual system) was weighted using sub-band division. In this method, the assessment valued of the new method accord with MOS (mean opinion score). In order to demonstrate the effectiveness and to verify the reliability of WVWPSNR method, a series of experiment for JPEG、 JPEG2000、 White Noise、 Gaussian Blur and FastFading images were made and tested. The results suggest that the proposed method can provide accurate measures of image quality under JPEG compression.%提出了一种新的基于小波变换和视觉加权的图像质量客观评价方法(WVWPSNR,wavelet and vision_weighted peak signal noise ratio).该方法将图像的DWT和HVS特性相结合,利用子图分解以及视觉加权处理实现图像质量的客观评价.对多幅分别经过JPEG压缩、JPEG2000压缩、White Noise(白噪声)、Gaussian Blur(高斯模糊)、FastFading降质的图像进行测试,实验结果表明该方法可靠、有效,与主观评价结果更接近.尤其是该方法对JPEG压缩图像的质量评价性能较好.