Conformer and Blind Noisy Students for Improved Image Quality Assessment

被引:7
|
作者
Conde, Marcos, V [1 ]
Burchi, Maxime [1 ]
Timofte, Radu [1 ]
机构
[1] Univ Wurzburg, Inst Comp Sci, Comp Vis Lab, Wurzburg, Germany
关键词
SIMILARITY; INDEX;
D O I
10.1109/CVPRW56347.2022.00108
中图分类号
TP301 [理论、方法];
学科分类号
081202 ;
摘要
Generative models for image restoration, enhancement, and generation have significantly improved the quality of the generated images. Surprisingly, these models produce more pleasant images to the human eye than other methods, yet, they may get a lower perceptual quality score using traditional perceptual quality metrics such as PSNR or SSIM. Therefore, it is necessary to develop a quantitative metric to reflect the performance of new algorithms, which should be well-aligned with the person's mean opinion score (MOS). Learning-based approaches for perceptual image quality assessment (IQA) usually require both the distorted and reference image for measuring the perceptual quality accurately. However, commonly only the distorted or generated image is available. In this work, we explore the performance of transformer-based full-reference IQA models. We also propose a method for IQA based on semi-supervised knowledge distillation from full-reference teacher models into blind student models using noisy pseudo-labeled data. Our approaches achieved competitive results on the NTIRE 2022 Perceptual Image Quality Assessment Challenge: our full-reference model was ranked 4th, and our blind noisy student was ranked 3rd among 70 participants, each in their respective track.
引用
收藏
页码:939 / 949
页数:11
相关论文
共 50 条
  • [1] HOMOGENEITY BASED BLIND NOISY IMAGE QUALITY ASSESSMENT
    Huang, Xiaotong
    Chen, Li
    Tian, Jing
    Zhang, Xiaolong
    Fu, Xiaowei
    [J]. 2013 IEEE INTERNATIONAL CONFERENCE ON SYSTEMS, MAN, AND CYBERNETICS (SMC 2013), 2013, : 2963 - 2967
  • [2] Blind noisy image quality assessment using block homogeneity
    Huang, Xiaotong
    Chen, Li
    Tian, Jing
    Zhang, Xiaolong
    Fu, Xiaowei
    [J]. COMPUTERS & ELECTRICAL ENGINEERING, 2014, 40 (03) : 796 - 807
  • [3] Blind Image Quality Metric for Blurry and Noisy Image
    Seghir, Zianou Ahmed
    Hachouf, Fella
    Morain-Nicolier, Frederic
    [J]. 2013 IEEE SECOND INTERNATIONAL CONFERENCE ON IMAGE INFORMATION PROCESSING (ICIIP), 2013, : 193 - 197
  • [4] Improved Noisy Image Quality Assessment Using Multilayer Neural Networks
    Rubel, Andrii
    Rubel, Oleksii
    Abramova, Victoriya
    Proskura, Galina
    Lukin, Vladimir
    [J]. 2019 IEEE 2ND UKRAINE CONFERENCE ON ELECTRICAL AND COMPUTER ENGINEERING (UKRCON-2019), 2019, : 1046 - 1051
  • [5] Blind Noisy Image Quality Assessment Using Sub-Band Kurtosis
    Deng, Chenwei
    Wang, Shuigen
    Bovik, Alan C.
    Huang, Guang-Bin
    Zhao, Baojun
    [J]. IEEE TRANSACTIONS ON CYBERNETICS, 2020, 50 (03) : 1146 - 1156
  • [6] An improved blind/referenceless image spatial quality evaluator algorithm for image quality assessment
    Li, Xuesong
    Pan, Jinfeng
    Shang, Jianrun
    Souri, Alireza
    Gao, Mingliang
    [J]. INTERNATIONAL JOURNAL OF COMPUTATIONAL SCIENCE AND ENGINEERING, 2024, 27 (01) : 48 - 56
  • [7] Local Homogeneity Combined with DCT Statistics to Blind Noisy Image Quality Assessment
    Yang, Lingxian
    Chen, Li
    Chen, Heping
    [J]. SIXTH INTERNATIONAL CONFERENCE ON GRAPHIC AND IMAGE PROCESSING (ICGIP 2014), 2015, 9443
  • [8] Kurtosis-based Blind Noisy Image Quality Assessment in Wavelet Domain
    Wang, Shuigen
    Deng, Chenwei
    Li, Cheng
    Liu, Xun
    Zhao, Baojun
    [J]. 2015 IEEE INTERNATIONAL CONFERENCE ON SYSTEMS, MAN, AND CYBERNETICS (SMC 2015): BIG DATA ANALYTICS FOR HUMAN-CENTRIC SYSTEMS, 2015, : 1557 - 1560
  • [9] Blind image quality assessment with improved natural scene statistics model
    Zhang, Yazhong
    Wu, Jinjian
    Xie, Xuemei
    Li, Leida
    Shi, Guangming
    [J]. DIGITAL SIGNAL PROCESSING, 2016, 57 : 56 - 65
  • [10] Blind image quality assessment
    Li, X
    [J]. 2002 INTERNATIONAL CONFERENCE ON IMAGE PROCESSING, VOL I, PROCEEDINGS, 2002, : 449 - 452