The week's pick
Reseach Article
Semi-supervised Learning for Image Quality Assessment Problem
| International Journal of Computer Applications |
| Foundation of Computer Science (FCS), NY, USA |
| Volume 186 - Number 9 |
| Year of Publication: 2024 |
| Authors: Tuan Linh Dang, Thuy Ha Hoang, Minh Hoang Cu, Duc Quang Nguyen, Huu Phuc Hoang |
10.5120/ijca2024923435
|
Tuan Linh Dang, Thuy Ha Hoang, Minh Hoang Cu, Duc Quang Nguyen, Huu Phuc Hoang . Semi-supervised Learning for Image Quality Assessment Problem. International Journal of Computer Applications. 186, 9 ( Feb 2024), 9-13. DOI=10.5120/ijca2024923435
Abstract
We live in the 21st century, a period of digital data explosion. Images are one example. Millions of photos are created yearly, so how can we evaluate their quality? In this article, we will introduce SSL algorithms to solve the problem of image quality assessment. We combined the KONIQ-10K and KADIS-700K datasets to create a new dataset and fix the image quality issues in the old datasets. We conducted comprehensive testing on the Vision Transformer in combination with 5 SSL algorithms, and the results we obtained were exceptional. Compared to ViT, ViT combined with the CRMatch algorithm gave outstanding results, with MAE reduced from 0.53 to 0.40.
References
- H.R.Sheikh, M. F. Sabir and A. C. Bovik, “A statistical evaluation of recent full reference image quality assessment algorithms,” IEEE Transactions on image processing, jourvol 15, number 11, pages 3440–3451, 2006
- N. Ponomarenko, V. Lukin, A. Zelensky, K. Egiazarian, M. Carli and F. Battisti, “Tid2008-a database for evaluation of full-reference visual quality assessment metrics,” Advances of modern radioelectronics, jourvol 10, number 4, pages 30–45, 2009.
- N. Ponomarenko, O. Ieremeiev, V. Lukin andothers, “Color image database tid2013: Peculiarities and preliminary results,” inEuropean workshop on visual information processing (EUVIP) IEEE, 2013, pages 106–111
- V. Hosu, H. Lin, T. Sziranyi and D. Saupe, “Koniq-10k: An ecologically valid database for deep learning of blind image quality assessment,” IEEE Transactions on Image Processing, jourvol 29, pages 4041–4056, 2020.
- KADID-10k: A Large-scale Artificially Distorted IQA Database, Hanhe Lin, Vlad Hosu, and Dietmar Saupe
- K. Sohn, D. Berthelot, N. Carlini andothers, “Fixmatch: Simplifying semisupervised learning with consistency and confidence,” Advances in neural information processing systems, jourvol 33, pages 596–608, 2020.
- Adamatch: A Unified approach to semi supervised learning and domain adaptation David Berthelot , Rebecca Roelofs , Kihyuk Sohn , Nicholas Carlini , Alex Kurakin
- CoMatch: Semi-supervised Learning with Contrastive Graph Regularization Junnan Li Caiming Xiong Steven C.H. Hoi
- FREEMATCH: SELF-ADAPTIVE THRESHOLDING FOR SEMI-SUPERVISED LEARNING Yidong Wang12, Hao Chen3 , Qiang Heng4, Wenxin Hou5, Yue Fan6, Zhen Wu7, Jindong Wang1, Marios Savvides3, Takahiro Shinozaki2, Bhiksha Raj3, Bernt Schiele6, Xing Xie1
- Revisiting Consistency Regularization for Semi-Supervised Learning Yue Fan Anna Kukleva Bernt Schiele
- Zheng, Mingkai, et al. ”Simmatch: Semi-supervised learning with similarity matching.” Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 2022.
- Su, Shaolin, et al. ”Blindly assess image quality in the wild guided by a self-adaptive hyper network.” Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 2020.
- Zhang, Weixia, et al. ”Blind image quality assessment using a deep bilinear convolutional neural network.” IEEE Transactions on Circuits and Systems for Video Technology 30.1 (2018): 36-47.
- Dosovitskiy, Alexey, et al. ”An image is worth 16x16 words: Transformers for image recognition at scale.” arXiv preprint arXiv:2010.11929 (2020).
- Re-IQA: Unsupervised Learning for Image Quality Assessment in theWild Avinab Saha Sandeep Mishra Alan C. Bovik Laboratory of Image and Video Engineering, The University of Texas at Austin
Index Terms
Keywords