摘要
图像在采集、传输或者显示的过程中都会导致图像失真,所以如何客观评价图像质量是数字图像/视频处理领域中极具研究价值的一门课题。客观图像质量评价方法分为三大类:全参考方法、半参考方法和无参考方法。论文选取了经典的2种全参考算法和5种具有代表性的无参考算法进行分析,并在四个公开的数据库上进行算法性能验证。
Therefore,how to evaluate the image quality objectively is an issue of great research value in the field of digital im. age/video processing in the process of image acquisition,transmission or display. The objective image quality evaluation methods are divided into three categories:full-reference image quality evaluation method,reduce-reference image quality evaluation meth. od and no-reference image quality evaluation method. In this paper,two classical full-reference algorithms and five representative no-reference algorithms are selected for analysis. The performance of the algorithm is verified in four public databases.
作者
孙翠霞
刘有耀
SUN Cuixia;LIU Youyao(School of Computer,Xi'an University of Posts and Telecommunications,Xi'an 710061)
出处
《计算机与数字工程》
2019年第9期2290-2294,共5页
Computer & Digital Engineering
关键词
全参考图像质量评价
无参考图像质量评价
图像质量评价数据库
算法性能
full-reference image quality assessment
no-reference image quality assessment
databases of image quality as. sessment
algorithm performance