中文摘要：

为了进一步提高全参考图像质量评价（IQA）方法的准确性和高效性,基于人类视觉系统（HVS）的低层次视觉感知机理,在利用梯度强度信息的同时还巧妙地利用了梯度方向信息,更好地表达了图像局部结构信息的失真。在LIVE2图像数据库上的实验结果表明：对不同类型的图像失真,联合利用梯度强度与方向信息的全参考IQA方法相比之前著名的基于HVS以及仅利用梯度强度信息的IQA方法具有更高的预测单调性和准确性,而且计算复杂度更低。

英文摘要：

In order to further improve the accuracy and efficiency of full reference image quality assessment（ IQA） methods,based on the low level visual perception mechanism of human visual system（ HVS）,this paper subtly utilizes the information of gradient direction while using the information of gradient strength,and thus expresses the distortion of local structural information of the images more effectively. The experiment results on LIVE2 image database show that,for different types of image distortions,the proposed joint gradient strength and gradient direction full reference IQA method achieves better prediction monotonicity and accuracy compared with prior famous HVS based IQA methods and the methods only utilizing gradient strength,and keeps lower computation complexity.