基于显著性图像边缘的全参考图像质量评价

Full reference image quality assessment based on the edge of saliency image

大部分图像质量评价算法仅研究图像的灰度值及信息内容,没有充分考虑人眼视觉特性对图像质量评价的影响,针对此问题提出了一种基于显著性图像边缘的全参考图像质量评价方法。利用显著性图像表征图像的视觉关注内容,然后提取显著性图像的边缘检测图,得到人眼关注的结构和边缘,最后计算参考图像与失真图像边缘检测图之间的汉明距离得到图像质量评价指标。实验结果表明本文提出的评价指标满足人眼视觉特性,对于各种失真类型有非常高的主观一致性,评价性能也优于很多其他指标。

Most of the image quality assessment algorithms, which only study the image gray value and the content of information , have not been fully taken into account the effect of the human visual system in the assessment of image quality. To solve this problem, a full reference image quality evaluation method based on the saliency image edge is proposed. Using the saliency image to represent the visual attention of the image, the salient image edge detection map is extracted to get the structure and the edge, and the Hamming distance of the edge detection map between the reference image and the distortion image edge is calculated to achieve the image quality assessment index. Experimental results show that the proposed assessment index can satisfy the human visual characteristics, and a very high subjective consistency can be realized for all kinds of distortion types. The performance of the proposed method is also superior to other indices.