基于多尺度拉东变换的图像感知质量评价

Quality estimation of multi-scale Radon transform for image visual

[目的]量化图像退化程度,提出了一种基于多尺度拉东变换的部分参考型图像质量评价方法。[方法]首先利用小波变换获得图像不同方向频率成分的小波子带,然后计算小波子带投影谱及投影谱的分形维数,并将这些分形维数组成向量作为图像的特征表示,最后计算原始图像与失真图像特征向量的1范数预测图像质量。[结果]该方法在LIVE标准图像库上进行测试,线性相关系数PLCC为0.949 1,等级相关系数SROCC为0.951 9,优于传统方法。[结论]该方法得到的客观质量评价与人的主观质量评价之间具有较好的一致性。

[Objective]To quantify image degradation degree,a reduced-reference image quality assessment metric based on multi-scale Radon transform was proposed in this paper.[Method]The orientation and frequency components of an image were first extracted by wavelet transform.Then Radon transform method was used so as to obtain projection spectrums on each sub-band.Next fractal dimensions was adopted to character each projection and concatenated as the image features.Finally L1 distance is used in order to pool the image features as the objective quality score.[Results]The proposed metric was tested on the LIVE database,experimental results have showed that the PLCC and SROCC were 0.9491,and 0.9519,respectively,which was much better than traditional methods.[Conclusion]The proposed metric has a good consistency with subjective perception.