电子学报 ›› 2021, Vol. 49 ›› Issue (11): 2261-2272.DOI: 10.12263/DZXB.20200780

• 综述评论 • 上一篇    下一篇

全参考图像质量评价回顾与展望

高敏娟1, 党宏社1, 魏立力2, 刘国军2, 张选德3   

  1. 1.陕西科技大学电气与控制工程学院,陕西 西安 710021
    2.宁夏大学数学统计学院,宁夏 银川 750021
    3.陕西科技大学电子信息与人工智能学院,陕西 西安 710021
  • 收稿日期:2020-07-27 修回日期:2021-03-09 出版日期:2021-11-25
    • 作者简介:
    • 高敏娟 女,1984年生,山西忻州人.博士研究生,主要研究方向为图像处理、图像质量评价.E-mail:gaominjuan1984@163.com
      党宏社 男,1962年生,陕西武功人.教授、博士生导师,主要研究方向为工业过程与优化、计算机控制和图像处理. E-mail:danghs@sust.edu.cn
      魏立力 男,1965年生,甘肃天水人.教授,主要研究方向为应用统计与数据分析. E-mail:liliwei@nxu.edu.cn
      刘国军 男,1978年生,宁夏吴忠人.教授,主要研究方向为小波分析、偏微分方程图像处理和图像质量评价. E-mail:liugj@nxu.edu.cn
      张选德(通讯作者) 男,1979年生,宁夏固原人.教授、博士生导师,主要研究方向为图像恢复、图像质量评价、稀疏表示和低秩逼近理论. E-mail:zhangxuande@sust.edu.cn
    • 基金资助:
    • 国家自然科学基金 (61871260); 陕西省自然科学基金 (2020JM-509)

Review and Prospect of Full Reference Image Quality Assessment

GAO Min-juan1, DANG Hong-she1, WEI Li-li2, LIU Guo-jun2, ZHANG Xuan-de3   

  1. 1.School of Electrical and Control Engineering, Shaanxi University of Science and Technology, Xi’an, Shaanxi 710021, China
    2.School of Mathematics and Statistics, Ningxia University, Yinchuan, Ningxia 750021, China
    3.School of Electronic Information and Artificial Intelligence, Shaanxi University of Science and Technology, Xi’an, Shaanxi 710021, China
  • Received:2020-07-27 Revised:2021-03-09 Online:2021-11-25 Published:2021-11-25
    • Supported by:
    • National Natural Science Foundation of China (61871260); Natural Science Foundation of Shaanxi Province, China (2020JM-509)

摘要:

全参考图像质量评价(Full Reference Image Quality Assessment, FR-IQA)是IQA领域广为研究的类型之一.本文回顾了FR-IQA的发展历程,对FR-IQA应用现状和通用FR-IQA问题的构建进行综述,以及对FR-IQA算法进行总结和梳理.并在此基础上,重点分析了现有研究中存在的问题,包括问题构建的合理性、建模的全面性问题、知识驱动与数据驱动结合的问题等.基于对主观评价过程的深入分析,结合现有研究存在的问题,探讨了主观评分采用模糊建模和知识与数据联合驱动构建算法两个可能的研究方向,以期对后续的研究者提供参考.

关键词: 图像质量评价, FR-IQA, 主观评分, 模糊建模, 知识与数据联合驱动

Abstract:

Full reference image quality assessment (FR-IQA) is one of the types widely studied in the field of IQA. This paper reviews the development of FR-IQA, summarizes the application status of FR-IQA and the construction of general FR-IQA problems, and summarizes and combs FR-IQA algorithms. And on this basis, it focuses on the problems existing in the existing research, including the rationality of the problem construction, the comprehensive problem of modeling, and the problem of the combination of knowledge-driven and data-driven. Finally, based on the in-depth analysis of the subjective evaluation process and the existing problems in the existing research, the two possible research directions of subjective scoring using fuzzy modeling and knowledge-data-driven construction algorithm are discussed, in order to provide reference for subsequent researchers.

Key words: image quality assessment, FR-IQA, subjective rating, fuzzy modeling, knowledge-data joint driven

中图分类号: