基于色差的均方误差与峰值信噪比评价彩色图像质量研究

Study on Color Image Quality Evaluation by MSE and PSNR Based on Color Difference

在图像质量评价模型中加入人眼视觉系统特性能够提高其评价性能.彩色图像色差模型S-CIELAB是在考虑到了人眼视觉系统的空间模糊特性基础上得到两幅图像的色差,该文在S- CIELAB色差模型的基础上,提出了色差均方误差(CD-MSE)和色差峰值信噪比(CD-PSNR)彩色图像质量客观评价参量.对JEPG2000、加白噪音和图像传输中产生位错的三类失真共180幅彩色图像进行了CD-PSNR参量计算,并与国际上已经给出的主观评价结果进行了比较.实验结果表明其客观评价结果与平均主观评价分数具有较好的一致性.

The performance of an image quality evaluation model can be improved by considering characteristics of the human visual system. The S-CIELAB model, which takes account of the spatial blurring by human visual system, is used to measure color difference between two color images. In this paper,it presents that the two parameters of objective evaluation of color image quality, color difference mean square error（CD-MSE） and color difference peak signal-to-noise ratio（CD-PSNR） based on the model of S-CIELAB. Objective experiments are conducted on three kinds of distorted images of JEPG2000, white Gaussian noise and transmission errors in JPEG2000 bitstream（total 180 images）, compared with subjective evaluation of mean opinion score（MOS）,and results show consistency between them.