基于边缘信息的图像质量评价研究  

Research on Image Quality Assessment Based on Edge Information

作　　者：王春哲[1] 李杰[2] 

机构地区：[1]长春理工大学电子信息工程学院,长春130022 [2]长春大学电子信息工程学院,长春130022

出　　处：《长春大学学报》2014年第10期1330-1333,共4页Journal of Changchun University

基　　金：吉林省教育厅科研项目(2013264);吉林省科技厅自然基金项目(201215107)

摘　　要：针对传统客观图像质量评价方法没有考虑人眼的视觉信息这一问题,本文提出了一种基于边缘信息的图像质量评价方法。该方法考虑到人眼对图像的边缘信息敏感,运用Prewitt算子提取参考图像与失真图像的边缘信息,对两图像的边缘信息做均方误差,来衡量失真图像的质量,实验结果表明,该方法比传统评价方法更接近于图像的实际质量。In view of the problem that the traditional objective image quality evaluation method does not take human eye visual informa- tion into account, this paper puts forward a kind of image quality evaluation method based on edge information, which considers that human eyes are sensitive to the edge information of the image, uses Prewitt operator to extract the edge information of the reference image and distorted image and calculates the mean square error of the edge information of the two images, to measure the quality of the distorted image. The experimental results show that the method is closer to the actual quality of the image than traditional evaluation methods.

关 键 词：图像边缘信息 均方误差 图像质量评价 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]