全参考图像质量指标评价分析  

Evaluation and analysis of full reference image quality indicators

作　　者：王成 刘坤 杜砾 WANG Cheng;LIU Kun;DU Li(School of Electronic Engineering,Xi’an Aeronautical Institute,Xi’an 710077,China)

机构地区：[1]西安航空学院电子工程学院,陕西西安710077

出　　处：《现代电子技术》2023年第21期39-43,共5页Modern Electronics Technique

摘　　要：图像是人类从外界获取信息的重要来源,通过客观的图像质量评价,能帮助人们在浏览图像时关注高质量的图像,提升感观体验。因此,如何对大量的图像质量进行客观而有效的评价,已成为图像信息处理领域的研究热点之一。根据对参考图像(即无失真的原始图像)的依赖程度,客观图像评价方法可分为全参考图像质量评价、部分参考图像质量评价和无参考图像质量评价。针对全参考图像质量评价方法进行分析,比较了常见的4种全参考图像质量评价指标,验证了这些评价指标分数与主观视觉评估相一致。The image is an important source for human beings to obtain information from the outside world.Objective image quality evaluation can help people focus on high⁃quality images and enhance their sensory experience when browsing images.Therefore,how to objectively and effectively evaluate the quality of a large number of images has become one of the research hotspots in the field of image information processing.According to the degree of dependence on reference images(i.e.,undistorted original images),objective image evaluation methods can be classified into full reference image quality evaluation,partial reference image quality evaluation,and non reference image quality evaluation.In this article,the full reference image quality evaluation method is analyzed,and four common full reference image quality evaluation indicators are compared.It is verified that these evaluation indicators have consistent scores with subjective visual evaluation.

关 键 词：图像质量评价 全参考 平均主观分数 人类视觉系统 结构相似性 多尺度结构相似性 

分 类 号：TN911.73-34[电子电信—通信与信息系统] TP391.4[电子电信—信息与通信工程]