江汉大学学报(自然科学版) ›› 2020, Vol. 48 ›› Issue (1): 91-96.doi: 10.16389/j.cnki.cn42-1737/n.2020.01.012

• 计算机科学技术 • 上一篇    

基于双向过滤法的块效应下无参考图像质量评价

马璐1,2,杨文飞3   

  1. 1. 宿州职业技术学院 计算机信息系,安徽 宿州 234000;2. 南京大学 电子科学与工程学院,江苏 南京 210023;3. 宿州学院 外国语学院,安徽 宿州 234000
  • 发布日期:2020-01-20
  • 作者简介:马璐(1979— ),男,讲师,硕士,南京大学访问学者,研究方向:图像处理、图像质量评价、三维立体图像。
  • 基金资助:
    安徽省教育厅高等学校自然科学重点项目(KJ2017A657);安徽省教育厅高校优秀青年骨干人才项目(gxgnfx2018112)

No-Reference Image Quality Assessment Under Blocking Artifacts Based on Bidirectional Filtration

MA Lu1,2,YANG Wenfei3   

  1. 1. Department of Information and Computer,Suzhou Vocational and Technical College,Suzhou 234000,Anhui,China;2. School of Electronic Science and Engineering,Nanjing University,Nanjing 210023,Jiangsu,China;3. School of Foreign Languages,Suzhou University,Suzhou 234000,Anhui,China
  • Published:2020-01-20

摘要: 针对块效应下无参考图像质量评价,提出了双向过滤法(bidirectional filtration),通过分辨图像真边界与假边界进行块效应下图像评价。该方法从两个不同的场域对图像边界进行逐层过滤,达到对真边界的有效定位,最后通过评价模型得出评价值。实验测试结果得出的数据与SSIM 模型相比有了一定的改进。

关键词: 双向过滤法, 图像质量评价, 图像边界

Abstract: To evaluate the image quality without reference under blocking effect,the author proposed a bidirectional filtration (BIF) method,which aims at providing a new idea for boundary selection,and resolved the image evaluation under blocking effect more effectively by distinguishing the true and false boundary of the image. The method filtered the image boundary layer by layer from two different fields to locate the true boundary effectively. Finally,the evaluation values were obtained through the evaluation model. Compared with SSIM model,the data obtained from the experimental results were improved to a certain extent. It is concluded that this method is more accurate in evaluating the no-reference image under blocking effect.

Key words: bidirectional filtration, image quality assessment, image boundary

中图分类号: