计算机工程与应用 ›› 2020, Vol. 56 ›› Issue (22): 191-198.DOI: 10.3778/j.issn.1002-8331.1910-0251

• 图形图像处理 • 上一篇    下一篇

基于MOS的图像质量评估系统

韩东旭,钟宝江   

  1. 苏州大学 计算机科学与技术学院,江苏 苏州 215000
  • 出版日期:2020-11-15 发布日期:2020-11-13

Image Quality Assessment System Based on MOS

HAN Dongxu, ZHONG Baojiang   

  1. School of Computer Science and Technology, Soochow University, Suzhou, Jiangsu 215000, China
  • Online:2020-11-15 Published:2020-11-13

摘要:

图像质量评估在图像处理技术的发展中起着关键的作用。针对当前图像质量主观评估过程中存在的操作复杂度高、执行效率低以及容易掺杂特定偏好等问题,研究并设计了一款图像质量评估系统。该系统通过隐藏待评估样本的标签和次序,从而消除了测评者可能存在的特定偏好。采用评估给分双确认的策略,并基于平均意见得分(Mean Opinion Score,MOS)对测评结果进行整理和分析,保证了主观评估的可靠性和执行效率。对现有的直线段检测算法进行评估,验证了系统的有效性。

关键词: 图像质量评估, 主观评估, 评估系统, 平均意见得分(MOS), 双确认评估, 标签隐藏, 次序隐藏

Abstract:

Image quality assessment plays an important role in the developments of various image processing techniques. The current subjective assessment of image quality has a complex evaluation process and thus is inefficient. Furthermore, the evaluation results could be easily effected by a possible bias of the evaluators. To solve the above-mentioned problems, a system is designed for conducting a reliable and effective subjective assessment. In this system, the label and order of each image sample are hidden, so that the possible bias of the evaluators can be effectively eliminated. This paper also adopts a strategy of double confirmation for each evaluation score, and analyzes the evaluation results by using the Mean Opinion Score(MOS). To verify the efficiency of the developed system, a subjective assessment of the state-of-the-art algorithms for line segment detection is conducted finally.

Key words: image quality assessment, subjective assessment, assessment system, Mean Opinion Score(MOS), double-confirmed assessment, label hiding, order hiding