一种基于颜色特征的图像质量评价算法  

Algorithm for Image Quality Assessment Based on Color Features

作　　者：高攀[1] 杨爽[1] 孟放[1] 

机构地区：[1]中国传媒大学信息工程学院,北京100024

出　　处：《电视技术》2012年第16期98-100,共3页Video Engineering

基　　金：国家广播电影电视总局科技项目:高清图像质量客观评测软件开发

摘　　要：颜色作为描述图像最直接有效的特征,是目前图像质量评价中备受关注的评价参数之一。提出一种基于颜色特征的全参考图像质量评价系统,该系统在结构相似性SSIM(Structural Similarity)理论的基础上,充分考虑色度信息对人眼视觉感知的影响,利用分块颜色矩,建立了一种新的相似度度量模型。实验结果表明,该模型获得的客观评价结果与主观评价值具有很高的一致性。As the most direct and efficient features to describe images, color features currently become one of the most significant parameters in image quality assessment. Based on the theory of structural similarity, a model of full-reference objective quality assessment system for image on the basis of color features is proposed,which takes the affect of chrominance on the human visual perception into consideration. After the theoretical analysis, a new similarity model using color moments of partitions is established in this paper. Experimental results show that the model has high consistency with the subjective results.

关 键 词：图像质量评价 SSIM 颜色特征 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]