Pith. sign in

REVIEW 4 major objections 6 minor 57 references

This challenge paper reports that four quality-assessment methods predict human perceptual scores on super-resolved images from modern GAN and diffusion models, all exceeding 0.9 on the ISRGen-QA test set.

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

A new SR image quality dataset focused on modern GAN and diffusion super-resolution outputs, plus benchmark results from four teams achieving SRCC above 0.90, is presented.

T0 review reviewed 2026-08-04 challenge →

load-bearing objection A useful new public dataset for SR-IQA on modern generative SR, but the SOTA claim and test results are weakened by an unspecified content split and no baselines. the 4 major comments →

arxiv 2509.06413 v1 pith:3DQFBSHP submitted 2025-09-08 cs.CV eess.IV

VQualA 2025 Challenge on Image Super-Resolution Generated Content Quality Assessment: Methods and Results

classification cs.CV eess.IV
keywords image super-resolutionquality assessmentgenerative modelsGAN artifactsdiffusion SRmean opinion scoreSRCCPLCC
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper presents the first benchmark challenge specifically for assessing the perceptual quality of images super-resolved by recent generative models — GANs, diffusion models, and transformer-based upscalers. It claims that all four submitted teams achieved overall scores above 0.9, meaning their predicted quality ratings closely track human mean opinion scores. This matters because older SR-IQA datasets contain few or no generative SR outputs, so existing metrics may not capture the artifacts of current super-resolution technology. The paper argues that blind, no-reference methods can succeed on this new benchmark, and the winning method needed only the SR image itself, not the high-resolution reference.

Core claim

The paper's central claim is that the ISRGen-QA dataset — 720 approximately 2K SR images from 15 recent super-resolution algorithms (ESRGAN, Real-ESRGAN, SR3, SwinIR, and others), rated by 21 valid subjects — supports reliable automatic quality prediction, and that the four challenge submissions achieve state-of-the-art correlation with human opinion. The best method (MICV) reached an overall score of 0.9638, combining SRCC 0.9588 and PLCC 0.9714. All teams' PLCC values exceeded their SRCC values, indicating the models capture linear relationships well but occasionally misorder closely rated images. The paper further describes each method: a hybrid ViT-CNN that uses only the SR image, a BLIP

What carries the argument

The central object is the ISRGen-QA dataset itself, built from 19 high-resolution references and 76 low-resolution images downsampled at four scales, then super-resolved by 15 recent GAN, diffusion, transformer, flow, and CNN methods to yield 720 SR images at roughly 2K resolution with MOS from 21 human raters. The evaluation machinery is the weighted score Score = 0.6×SRCC + 0.4×PLCC, which jointly rewards rank-order and linear agreement with human scores. Each submitted method is a distinct learned regressor mapping SR images (and sometimes the LR reference) to quality scores, trained on the 576-image development split.

Load-bearing premise

The train/validation/test split of the 720 SR images is not shown to be disjoint at the level of the 19 high-resolution reference images, so artifacts specific to a particular reference could in principle appear in both training and test and inflate the reported correlations.

What would settle it

Regroup the 720 SR images by their high-resolution parent image and split so that all SR images from any single HR reference stay in exactly one split; retrain the submitted models on the content-disjoint training fold. If the test correlations drop substantially below the reported 0.9, the original numbers were inflated by reference leakage. A cheaper check is to verify whether any same-HR image pairs straddle the current train/test boundary.

Watch this falsifier. Get emailed when new claim-graph text bears on it.

If this is right

  • If the central claim holds, blind no-reference quality metrics can reliably grade images produced by modern generative super-resolution, removing the need for high-resolution references in quality monitoring.
  • The benchmark provides a reproducible test bed for comparing SR-IQA methods on the artifact types actual users encounter today, such as hallucinated textures and unnatural reconstruction patterns.
  • The patterns in the results — high PLCC with slightly lower SRCC — suggest that future methods should focus on fine-grained rank ordering among similar-quality images, where current models are weakest.
  • Because the winning method ignores LR and HR references, lightweight SR-only quality models could be deployed in real-world pipelines without extra reference data.
  • The dataset's coverage of four upscaling factors and 15 algorithms gives a concrete target for improving generalization across SR methods.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The reported numbers rest on a train/test split that is not shown to be disjoint at the level of the 19 high-resolution reference images; if the same reference contributes both training and test SR images, models could memorize reference-specific artifacts and inflate correlations.
  • With only 72 test images and one MOS value per image, the 0.9+ correlations may carry wide confidence intervals; a content-disjoint split or repeated resampling would reveal how stable the ranking is.
  • The consistent PLCC-over-SRCC gap across all four teams hints that the test set may contain distinctive quality clusters rather than a smooth spectrum, a pattern worth checking on other generative SR datasets.
  • A natural extension is to test whether these models transfer across SR datasets with different generative backbones (e.g., from GAN-heavy to diffusion-heavy outputs), since all four teams trained and tested on the same small dataset.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit.

Referee Report

4 major / 6 minor

Summary. This paper reports the organization and results of the VQualA 2025 ISRGC-Q challenge on quality assessment of super-resolved (SR) images produced by modern generative models. It introduces the ISRGen-QA dataset (720 SR images from 15 SR algorithms, derived from 19 HR and 76 LR references, with MOS from 21 valid participants), defines an evaluation protocol based on a 0.6·SRCC + 0.4·PLCC score, and describes the four participating methods. All four teams report overall scores above 0.9 on a 72-image test set, and the abstract claims these results demonstrate state-of-the-art performance. The paper also provides architecture and training details for each method and makes the dataset publicly available.

Significance. If the reported numbers are valid, the challenge demonstrates that learned models can predict human perceptual quality ratings on modern GAN- and diffusion-based SR content with high correlation. The public ISRGen-QA dataset is a potentially valuable resource, as existing SR-IQA datasets predate the current generation of generative SR methods. The paper also provides detailed method descriptions and a reproducible challenge protocol, which are strengths. However, the central SOTA claim is not yet supported because no comparison with existing NR-IQA or SR-IQA baselines is provided, and the 72-image test set without uncertainty quantification makes the absolute scores difficult to interpret.

major comments (4)
  1. [Section 2.1] The dataset split is described only as 'training (576 images, 80%), validation (72 images, 10%), and test (72 images, 10%)' without specifying whether the split is at the SR-image level or at the level of the 19 HR / 76 LR content sources. Because each content source yields multiple SR outputs (different algorithms and scales), an SR-level split is likely to place different versions of the same underlying content in training and test. This would let reference-based methods (2077 Agent, Sec. 4.4, uses HR; ydy, Sec. 4.2, uses LR) exploit content correspondences, and would allow even SR-only models to memorize content-specific artifact patterns. This is load-bearing for the validity of the reported test correlations. Please state the split granularity; if it is SR-level, re-split by content and re-report all results.
  2. [Abstract / Section 3.1] The abstract and Section 3.1 describe the submitted methods as 'state-of-the-art' with 'exceptional performance' based on overall scores exceeding 0.9. No existing NR-IQA or SR-IQA baselines are evaluated on the same test set. Without comparisons to, e.g., MUSIQ, NIQE, BRISQUE, or SR-specific metrics, the absolute SRCC/PLCC values are uncalibrated and cannot support a SOTA claim. Please add a baseline comparison on the same test split.
  3. [Section 3.1] All conclusions in Section 3.1 are based on 72 test images with no confidence intervals or significance tests. The differences among teams (e.g., SRCC 0.9588 vs. 0.9277) are small relative to the sampling variability of Spearman correlation at n=72. The claims that PLCC consistently exceeds SRCC and the ranking of methods are not statistically established. Report bootstrap confidence intervals or significance tests, and temper the 'consistent pattern' conclusions accordingly.
  4. [Section 2.2] The evaluation protocol defines PLCC but does not state whether the Pearson correlation is computed on raw predicted scores or after a nonlinear monotonic mapping (e.g., a logistic fit), which is standard in IQA. This affects the reported PLCC values and comparability with prior SR-IQA work. Please specify the exact computation and whether any nonlinear regression was applied before computing PLCC.
minor comments (6)
  1. [Section 2.3.2] There is an inconsistency: Section 2.3.2 says '5 teams submitted their final results' but the abstract and Section 3 say only 4 teams submitted valid solutions and fact sheets. Please clarify the count.
  2. [Section 4.1] The 'Testing Details' paragraph contains a duplicated sentence: 'Subsequently, center cropping is performed to extract 448×448 resolution images...' appears twice, with the second occurrence being incomplete. Please fix the editing error.
  3. [Section 2.1] The description 'anomaly filtering yielding valid scores from 21 participants' lacks detail. Specify how anomalies were detected and how many participants were excluded.
  4. [Eq. (1)] The word 'reweighting' is misleading; Eq. (1) is a weighted combination of SRCC and PLCC. Also clarify whether PLCC is computed after any nonlinear fitting, as noted in the major comments.
  5. [Table 1] The caption notes that GFlops are 'calculated relative to Input Size' but the precise meaning is unclear. State whether this is the multiply-accumulate count for the listed input resolution and how it was computed.
  6. [Figure 2] The fourth-order polynomial fitting is mentioned only for visualization. If this fit is also used in PLCC computation, it must be described in Section 2.2; otherwise, clarify that it is only for plotting.

Circularity Check

0 steps flagged

No circular derivation: test MOS are held out and the reported correlations come from external team submissions on a fixed metric.

full rationale

The paper is a challenge report; its central claim is that four submitted methods achieve SRCC/PLCC >0.9 on the held-out test split of ISRGen-QA (Sec. 3.1, Table 1). The evaluation chain is not circular: participants trained on 576 images with MOS and were evaluated on 72 test images whose MOS were withheld during development (Sec. 2.3). The ranking metric (Eq. 1) is a fixed linear combination of SRCC and PLCC; it does not encode the test outcomes. Each team's method description (Secs. 4.1-4.4) is an external architecture trained on the released training partition; no method fits parameters to test MOS. The paper's references include self-citations to challenge and team papers ([14], [17], [29], [49], [27]), but these are descriptive citations to the methods being summarized, not load-bearing justifications of the empirical results. The dataset and challenge are organized by the same group that reports the results, but the scores are computed on an out-of-sample test set by the organizers, so this is standard benchmark practice rather than a derivation-from-input. The one methodological omission - Sec. 2.1 does not state whether the 576/72/72 split is content-disjoint given only 19 HR references and 76 LR references - is a potential data-leakage risk for the validity of the >0.9 numbers, but it is not a circularity: no equation reduces to itself and no fitted parameter is renamed a prediction. The absence of a baseline NR-IQA comparison in Sec. 3 is a correctness/context concern, not a circular step. No circular steps found.

Axiom & Free-Parameter Ledger

0 free parameters · 4 axioms · 0 invented entities

The paper introduces no new entities and fits no free parameters to the test set. Its central claim rests on domain assumptions about the reliability of the MOS labels, the validity of the data split, the choice of evaluation metrics, and the representativeness of the dataset content. These are all reasonable for a challenge report but are not independently verified in the paper.

axioms (4)
  • domain assumption MOS from 21 subjects after anomaly filtering is a reliable ground truth for perceptual quality
    Section 2.1: the dataset's quality labels rest on a small subject pool; no inter-rater agreement or confidence bounds are reported.
  • domain assumption The train/validation/test split is content-disjoint
    Section 2.1: 720 SR images from 19 HR references are split 80/10/10, but the split granularity is not specified; leakage would inflate test correlations.
  • domain assumption SRCC and PLCC with weights 0.6/0.4 are the appropriate evaluation metrics
    Section 2.2, Eq. 1: this is a contest design choice, presented without justification as the correct IQA evaluation.
  • domain assumption The 15 SR algorithms and the DIV2K references represent modern SR outputs
    Section 2.1: the selection of algorithms and content could bias the benchmark, and no evidence is given that they are representative.

reviewed 2026-08-04 · how reviews work

0 comments
Cite this review

Pith. "Pith review of VQualA 2025 Challenge on Image Super-Resolution Generated Content Quality Assessment: Methods and Results." pith.science (2026). https://pith.science/paper/3DQFBSHP

@misc{pith2026250906413,
  author       = {Pith},
  title        = {Pith review of: VQualA 2025 Challenge on Image Super-Resolution Generated Content Quality Assessment: Methods and Results},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/3DQFBSHP}},
  note         = {Machine review of arXiv:2509.06413}
}
Share X Bluesky LinkedIn Reddit HN
read the original abstract

This paper presents the ISRGC-Q Challenge, built upon the Image Super-Resolution Generated Content Quality Assessment (ISRGen-QA) dataset, and organized as part of the Visual Quality Assessment (VQualA) Competition at the ICCV 2025 Workshops. Unlike existing Super-Resolution Image Quality Assessment (SR-IQA) datasets, ISRGen-QA places a greater emphasis on SR images generated by the latest generative approaches, including Generative Adversarial Networks (GANs) and diffusion models. The primary goal of this challenge is to analyze the unique artifacts introduced by modern super-resolution techniques and to evaluate their perceptual quality effectively. A total of 108 participants registered for the challenge, with 4 teams submitting valid solutions and fact sheets for the final testing phase. These submissions demonstrated state-of-the-art (SOTA) performance on the ISRGen-QA dataset. The project is publicly available at: https://github.com/Lighting-YXLI/ISRGen-QA.

Figures

Figures reproduced from arXiv: 2509.06413 by Baoquan Zhao, Chenqi Zhang, Chris Wei Zhou, Chuanbiao Song, Dongyang Yan, Fan Yang, Guanghui Yue, Hadi Amirpour, Hang Ouyang, Huijia Zhu, Jun Lan, Junlin Chen, Kaixin Deng, Shishun Tian, Shuo Xing, Tianxin Xiao, Weide Liu, Weiqiang Wang, Wei Sun, Weixia Zhang, Xiaoshuai Hao, Xiaoyan Sun, Xiaoyuan Yang, Xin Li, Xinyu Li, Yixiao Li, Zhaowang Wu, Zhengzhong Tu, Zhihua Wang, Zhizun Luo, Zhuohang Shi.

Figure 1
Figure 1. Figure 1: The performance of the methods submitted by different [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Scatter plots of predicted scores versus MOS for all participating teams on the testing set. The curves are obtained by a fourth [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Model architecture of the proposed Hybrid Vision Trans [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Model architecture of the proposed Cross-Covariance [PITH_FULL_IMAGE:figures/full_fig_p005_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Model framework of the proposed BLIP-2 Assisted [PITH_FULL_IMAGE:figures/full_fig_p005_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Model framework of the proposed SRM: Semantic Re [PITH_FULL_IMAGE:figures/full_fig_p005_6.png] view at source ↗
Figure 8
Figure 8. Figure 8: The diagram of the proposed Blind Super-resolution [PITH_FULL_IMAGE:figures/full_fig_p006_8.png] view at source ↗
Figure 9
Figure 9. Figure 9: UltraR-IQA Network Architecture: Inputting SR/HR [PITH_FULL_IMAGE:figures/full_fig_p007_9.png] view at source ↗
Figure 10
Figure 10. Figure 10: FPN Module Architecture: Taking multi-level Con [PITH_FULL_IMAGE:figures/full_fig_p007_10.png] view at source ↗
Figure 11
Figure 11. Figure 11: Difference Contrast Network Architecture: Inputting [PITH_FULL_IMAGE:figures/full_fig_p007_11.png] view at source ↗
Figure 12
Figure 12. Figure 12: Multi-Scale Feature Fusion Architecture: Employing [PITH_FULL_IMAGE:figures/full_fig_p008_12.png] view at source ↗

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

57 extracted references · 55 canonical work pages

  1. [1]

    NTIRE 2017 Chal- lenge on Single Image Super-Resolution: Dataset and Study

    Eirikur Agustsson and Radu Timofte. NTIRE 2017 Chal- lenge on Single Image Super-Resolution: Dataset and Study. InProceedings of the IEEE Conference on Computer Vision and Pattern Recognition Workshops, 2017. 2

  2. [2]

    VQualA 2025 Chal- lenge on GenAI-Bench AIGC Video Quality Assessment: Methods and Results

    Ying Chen, Huasheng Wang, Pengxiang Xiao, Yukang Ding, Enpeng Liu, Chris Wei Zhou, and et al. VQualA 2025 Chal- lenge on GenAI-Bench AIGC Video Quality Assessment: Methods and Results. InProceedings of the IEEE/CVF Con- ference on Computer Vision (ICCV) Workshops, pages 1–11,

  3. [3]

    Second-Order Attention Network for Single Image Super-Resolution

    Tao Dai, Jianrui Cai, Yongbing Zhang, Shu-Tao Xia, and Lei Zhang. Second-Order Attention Network for Single Image Super-Resolution. InProceedings of the IEEE/CVF Con- ference on Computer Vision and Pattern Recognition, pages 11065–11074, 2019. 2

  4. [4]

    Implicit Diffusion Models for Contin- uous Super-Resolution

    Sicheng Gao, Xuhui Liu, Bohan Zeng, Sheng Xu, Yan- jing Li, Xiaoyan Luo, Jianzhuang Liu, Xiantong Zhen, and Baochang Zhang. Implicit Diffusion Models for Contin- uous Super-Resolution. InProceedings of the IEEE/CVF Computer Vision and Pattern Recognition Conference, pages 10021–10030, 2023. 2

  5. [5]

    A deep convolutional network for medical image super- resolution

    Yunxing Gao, Hengjian Li, Jiwen Dong, and Guang Feng. A deep convolutional network for medical image super- resolution. In2017 Chinese Automation Congress (CAC), pages 5310–5315, 2017. 2

  6. [6]

    VQualA 2025 Document Image Quality Assessment Challenge

    Fan Huang, Xiongkuo Min, Zhichao Ma, Xiaohong Liu, Chris Wei Zhou, Guangtao Zhai, and et al. VQualA 2025 Document Image Quality Assessment Challenge. InPro- ceedings of the IEEE/CVF Conference on Computer Vision (ICCV) Workshops, pages 1–8. 2

  7. [7]

    Single Image Super- Resolution Quality Assessment: A Real-World Dataset, Sub- jective Studies, and an Objective Metric.IEEE Transactions on Image Processing, 31:2279–2294, 2022

    Qiuping Jiang, Zhentao Liu, Ke Gu, Feng Shao, Xinfeng Zhang, Hantao Liu, and Weisi Lin. Single Image Super- Resolution Quality Assessment: A Real-World Dataset, Sub- jective Studies, and an Objective Metric.IEEE Transactions on Image Processing, 31:2279–2294, 2022. 1, 2

  8. [8]

    Deeply-Recursive Convolutional Network for Image Super- Resolution

    Jiwon Kim, Jung Kwon Lee, and Kyoung Mu Lee. Deeply-Recursive Convolutional Network for Image Super- Resolution. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 1637– 1645, 2016. 2

  9. [9]

    Ac- curate Image Super-Resolution Using Very Deep Convolu- tional Networks

    Jiwon Kim, Jung Kwon Lee, and Kyoung Mu Lee. Ac- curate Image Super-Resolution Using Very Deep Convolu- tional Networks. InProceedings of the IEEE/CVF Confer- ence on Computer Vision and Pattern Recognition, pages 1646–1654, 2016. 2

  10. [10]

    Photo- Realistic Single Image Super-Resolution Using a Generative Adversarial Network

    Christian Ledig, Lucas Theis, Ferenc Husz´ar, Jose Caballero, Andrew Cunningham, Alejandro Acosta, Andrew Aitken, Alykhan Tejani, Johannes Totz, Zehan Wang, et al. Photo- Realistic Single Image Super-Resolution Using a Generative Adversarial Network. InProceedings of the IEEE/CVF Con- ference on Computer Vision and Pattern Recognition, pages 4681–4690, 2017. 2

  11. [11]

    SeD: Semantic-Aware Discriminator for Image Super-Resolution

    Bingchen Li, Xin Li, Hanxin Zhu, Yeying Jin, Ruoyu Feng, Zhizheng Zhang, and Zhibo Chen. SeD: Semantic-Aware Discriminator for Image Super-Resolution. InProceedings of the IEEE/CVF Conference on Computer Vision and Pat- tern Recognition, pages 25784–25795, 2024. 2

  12. [12]

    VQualA 2025 Challenge on Engage- ment Prediction for Short Videos: Methods and Results

    Dasong Li, Sizhuo Ma, Hang Hua, Wenjie Li, Jian Wang, Chris Wei Zhou, Fengbin Guan, Xin Li, Zihao Yu, Yiting Lu, Ru-Ling Liao, Yan Ye, Zhibo Chen, Wei Sun, Linhan Cao, Yuqin Cao, Weixia Zhang, Wen Wen, Kaiwei Zhang, Zijian Chen, Fangfang Lu, Xiongkuo Min, Guangtao Zhai, Erjia Xiao, Lingfeng Zhang, Zhenjie Su, Hao Cheng, Yu Liu, Renjing Xu, Long Chen, Xiao...

  13. [13]

    SRDiff: Single image super-resolution with diffusion probabilistic models

    Haoying Li, Yifan Yang, Meng Chang, Shiqi Chen, Huajun Feng, Zhihai Xu, Qi Li, and Yueting Chen. SRDiff: Single image super-resolution with diffusion probabilistic models. Neurocomputing, 479:47–59, 2022. 2

  14. [14]

    Hybrid Vision Transformer and Convolutional Neural Network for Super- Resolution Image Quality Assessment

    Xinyu Li, Chuanbiao Song, Chenqi Zhang, Jun Lan, Hui- jia Zhu, Weiqiang Wang, and Xiaoyan Sun. Hybrid Vision Transformer and Convolutional Neural Network for Super- Resolution Image Quality Assessment. InProceedings of the IEEE/CVF Conference on Computer Vision (ICCV) Work- shops, 2025. 4

  15. [15]

    DifIISR: A Diffusion Model with Gradient Guidance for Infrared Im- age Super-Resolution

    Xingyuan Li, Zirui Wang, Yang Zou, Zhixin Chen, Jun Ma, Zhiying Jiang, Long Ma, and Jinyuan Liu. DifIISR: A Diffusion Model with Gradient Guidance for Infrared Im- age Super-Resolution. InProceedings of the IEEE/CVF Computer Vision and Pattern Recognition Conference, pages 7534–7544, 2025. 2

  16. [16]

    Deep bi-directional attention network for image super- resolution quality assessment

    Yixiao Li, Xiaoyuan Yang, Jun Fu, Guanghui Yue, and Wei Zhou. Deep bi-directional attention network for image super- resolution quality assessment. In2024 IEEE International Conference on Multimedia and Expo (ICME), pages 1–6. IEEE, 2024. 1

  17. [17]

    VQualA 2025 Challenge on Image Super-Resolution Generated Con- tent Quality Assessment: Methods and Results

    Yixiao Li, Xin Li, Chris Wei Zhou, Shuo Xing, Hadi Amir- pour, Xiaoshuai Hao, Guanghui Yue, Baoquan Zhao, Weide Liu, Xiaoyuan Yang, Zhengzhong Tu, and et al. VQualA 2025 Challenge on Image Super-Resolution Generated Con- tent Quality Assessment: Methods and Results. InPro- ceedings of the IEEE/CVF Conference on Computer Vision (ICCV) Workshops, pages 1–10...

  18. [18]

    SwinIR: Image Restoration Using Swin Transformer

    Jingyun Liang, Jiezhang Cao, Guolei Sun, Kai Zhang, Luc Van Gool, and Radu Timofte. SwinIR: Image Restoration Using Swin Transformer. InProceedings of the IEEE/CVF International Conference on Computer Vision, pages 1833– 1844, 2021. 2

  19. [19]

    Enhanced Deep Residual Networks for Single Image Super-Resolution

    Bee Lim, Sanghyun Son, Heewon Kim, Seungjun Nah, and Kyoung Mu Lee. Enhanced Deep Residual Networks for Single Image Super-Resolution. InProceedings of the IEEE Conference on Computer Vision and Pattern Recognition Workshops, pages 136–144, 2017. 2

  20. [20]

    CD- Former: When Degradation Prediction Embraces Diffusion Model for Blind Image Super-Resolution

    Qingguo Liu, Chenyi Zhuang, Pan Gao, and Jie Qin. CD- Former: When Degradation Prediction Embraces Diffusion Model for Blind Image Super-Resolution. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 7455–7464, 2024. 2

  21. [21]

    SGDR: Stochastic Gra- dient Descent with Warm Restarts

    Ilya Loshchilov and Frank Hutter. SGDR: Stochastic Gra- dient Descent with Warm Restarts. InInternational Confer- ence on Learning Representations, 2017. 6

  22. [22]

    Decoupled Weight De- cay Regularization

    Ilya Loshchilov and Frank Hutter. Decoupled Weight De- cay Regularization. InInternational Conference on Learning Representations, 2019. 6

  23. [23]

    Learning a no-reference quality metric for single-image super-resolution.Computer Vision and Image Understanding, 158:1–16, 2017

    Chao Ma, Chih-Yuan Yang, Xiaokang Yang, and Ming- Hsuan Yang. Learning a no-reference quality metric for single-image super-resolution.Computer Vision and Image Understanding, 158:1–16, 2017. 1

  24. [24]

    VQualA 2025 Challenge on Face Image Quality Assessment: Methods and Results

    Sizhuo Ma, Wei-Ting Chen, Qiang Gao, Jian Wang, Chris Wei Zhou, Wei Sun, Weixia Zhang, Linhan Cao, Jun Jia, Xiangyang Zhu, Dandan Zhu, Xiongkuo Min, Guang- tao Zhai, Baoying Chen, Xiongwei Xiao, Jishen Zeng, Wei Wu, Tiexuan Lou, Yuchen Tan, Chunyi Song, Zhiwei Xu, MohammadAli Hamidi, Hadi Amirpour, Mingyin Bai, Ji- awang Du, Zhenyu Jiang, Zilong Lu, Zigua...

  25. [25]

    Dynamic Attention- Guided Diffusion for Image Super-Resolution

    Brian B Moser, Stanislav Frolov, Federico Raue, Sebas- tian Palacio, and Andreas Dengel. Dynamic Attention- Guided Diffusion for Image Super-Resolution. In2025 IEEE/CVF Winter Conference on Applications of Computer Vision (WACV), pages 451–460, 2025. 2

  26. [26]

    Image Super- Resolution via Iterative Refinement.IEEE Transactions on Pattern Analysis and Machine Intelligence, 45(4):4713– 4726, 2022

    Chitwan Saharia, Jonathan Ho, William Chan, Tim Sali- mans, David J Fleet, and Mohammad Norouzi. Image Super- Resolution via Iterative Refinement.IEEE Transactions on Pattern Analysis and Machine Intelligence, 45(4):4713– 4726, 2022. 2

  27. [27]

    Ultra-High- Resolution Image Quality Assessment

    Zhuohang Shi, Zhizun Luo, Hang Ouyang, Tianxin Xiao, Fan Yang, Zhaowang Wu, and Kaixin Deng. Ultra-High- Resolution Image Quality Assessment. InProceedings of the IEEE/CVF Conference on Computer Vision (ICCV) Work- shops, 2025. 7

  28. [28]

    MDID: A multi- ply distorted image database for image quality assessment

    Wen Sun, Fei Zhou, and Qingmin Liao. MDID: A multi- ply distorted image database for image quality assessment. Pattern Recognition, 61:153–168, 2017. 1

  29. [29]

    BLIP-2 Assisted Residual- Guided Quality Assessment for Super-Resolution Images

    Shishun Tian and Dongyang Yan. BLIP-2 Assisted Residual- Guided Quality Assessment for Super-Resolution Images. In Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV) Workshops, 2025. 5

  30. [30]

    Boost- ing Flow-based Generative Super-Resolution Models via Learned Prior

    Li-Yuan Tsao, Yi-Chen Lo, Chia-Che Chang, Hao-Wei Chen, Roy Tseng, Chien Feng, and Chun-Yi Lee. Boost- ing Flow-based Generative Super-Resolution Models via Learned Prior. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 26005– 26015, 2024. 2

  31. [31]

    SigLIP 2: Multilingual Vision- Language Encoders with Improved Semantic Understanding, Localization, and Dense Features, 2025

    Michael Tschannen, Alexey Gritsenko, Xiao Wang, Muham- mad Ferjad Naeem, Ibrahim Alabdulmohsin, Nikhil Parthasarathy, Talfan Evans, Lucas Beyer, Ye Xia, Basil Mustafa, Olivier H ´enaff, Jeremiah Harmsen, Andreas Steiner, and Xiaohua Zhai. SigLIP 2: Multilingual Vision- Language Encoders with Improved Semantic Understanding, Localization, and Dense Feature...

  32. [32]

    Super-Resolution Quality Assessment: Subjective Evaluation Database and Quality Index Based on Perceptual Structure Measurement

    Wenfei Wan, Jinjian Wu, Guangming Shi, Yongbo Li, and Weisheng Dong. Super-Resolution Quality Assessment: Subjective Evaluation Database and Quality Index Based on Perceptual Structure Measurement. In2018 IEEE Interna- tional Conference on Multimedia and Expo (ICME), pages 1–6. IEEE, 2018. 1, 2

  33. [33]

    SAM- DiffSR: Structure-Modulated Diffusion Model for Image Super-Resolution.arXiv preprint arXiv:2402.17133, 2024

    Chengcheng Wang, Zhiwei Hao, Yehui Tang, Jianyuan Guo, Yujie Yang, Kai Han, and Yunhe Wang. SAM- DiffSR: Structure-Modulated Diffusion Model for Image Super-Resolution.arXiv preprint arXiv:2402.17133, 2024. 2

  34. [34]

    ESRGAN: En- hanced Super-Resolution Generative Adversarial Networks

    Xintao Wang, Ke Yu, Shixiang Wu, Jinjin Gu, Yihao Liu, Chao Dong, Yu Qiao, and Chen Change Loy. ESRGAN: En- hanced Super-Resolution Generative Adversarial Networks. InProceedings of the European Conference on Computer Vi- sion Workshops, pages 0–0, 2018. 2

  35. [35]

    Real-ESRGAN: Training Real-World Blind Super- Resolution With Pure Synthetic Data

    Xintao Wang, Liangbin Xie, Chao Dong, and Ying Shan. Real-ESRGAN: Training Real-World Blind Super- Resolution With Pure Synthetic Data. InProceedings of the IEEE/CVF Computer Vision and Pattern Recognition Con- ference, pages 1905–1914, 2021. 2

  36. [36]

    Attention

    Yan Wang, Yi Liu, Shijie Zhao, Junlin Li, and Li Zhang. CAMixerSR: Only Details Need More “Attention”. InPro- ceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 25837–25846, 2024. 2

  37. [37]

    Deep Networks for Image Super-Resolution With Sparse Prior

    Zhaowen Wang, Ding Liu, Jianchao Yang, Wei Han, and Thomas Huang. Deep Networks for Image Super-Resolution With Sparse Prior. InProceedings of the IEEE International Conference on Computer Vision, pages 370–378, 2015. 2

  38. [38]

    Super-Resolution Neural Op- erator

    Min Wei and Xuesong Zhang. Super-Resolution Neural Op- erator. InProceedings of the IEEE/CVF Conference on Com- puter Vision and Pattern Recognition, pages 18247–18256,

  39. [39]

    SeeSR: Towards Semantics-Aware Real-World Image Super-Resolution

    Rongyuan Wu, Tao Yang, Lingchen Sun, Zhengqiang Zhang, Shuai Li, and Lei Zhang. SeeSR: Towards Semantics-Aware Real-World Image Super-Resolution. InProceedings of the IEEE/CVF Computer Vision and Pattern Recognition Con- ference, pages 25456–25467, 2024. 2

  40. [40]

    Diffusion Prior Interpolation for Flexibil- ity Real-World Face Super-Resolution

    Jiarui Yang, Tao Dai, Yufei Zhu, Naiqi Li, Jinmin Li, and Shu-Tao Xia. Diffusion Prior Interpolation for Flexibil- ity Real-World Face Super-Resolution. InProceedings of the AAAI Conference on Artificial Intelligence, pages 9211– 9219, 2025. 2

  41. [41]

    Objective Quality Assessment of Interpolated Natural Im- ages.IEEE Transactions on Image Processing, 24(11):4651– 4663, 2015

    Hojatollah Yeganeh, Mohammad Rostami, and Zhou Wang. Objective Quality Assessment of Interpolated Natural Im- ages.IEEE Transactions on Image Processing, 24(11):4651– 4663, 2015. 1, 2

  42. [42]

    Scaling Up to Excellence: Practicing Model Scaling for Photo-Realistic Image Restoration In the Wild

    Fanghua Yu, Jinjin Gu, Zheyuan Li, Jinfan Hu, Xiangtao Kong, Xintao Wang, Jingwen He, Yu Qiao, and Chao Dong. Scaling Up to Excellence: Practicing Model Scaling for Photo-Realistic Image Restoration In the Wild. InProceed- ings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 25669–25680, 2024. 2

  43. [43]

    Deep Unfold- ing Network for Image Super-Resolution

    Kai Zhang, Luc Van Gool, and Radu Timofte. Deep Unfold- ing Network for Image Super-Resolution. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 3217–3226, 2020. 2

  44. [44]

    Designing a Practical Degradation Model for Deep Blind Image Super-Resolution

    Kai Zhang, Jingyun Liang, Luc Van Gool, and Radu Timo- fte. Designing a Practical Degradation Model for Deep Blind Image Super-Resolution. InProceedings of the IEEE/CVF Computer Vision and Pattern Recognition Conference, pages 4791–4800, 2021. 2

  45. [45]

    Transcending the Limit of Local Window: Ad- vanced Super-Resolution Transformer with Adaptive Token Dictionary

    Leheng Zhang, Yawei Li, Xingyu Zhou, Xiaorui Zhao, and Shuhang Gu. Transcending the Limit of Local Window: Ad- vanced Super-Resolution Transformer with Adaptive Token Dictionary. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 2856– 2865, 2024. 2

  46. [46]

    RankSRGAN: Generative Adversarial Networks With Ranker for Image Super-Resolution

    Wenlong Zhang, Yihao Liu, Chao Dong, and Yu Qiao. RankSRGAN: Generative Adversarial Networks With Ranker for Image Super-Resolution. InProceedings of the IEEE/CVF International Conference on Computer Vision, pages 3096–3105, 2019. 2

  47. [47]

    Uncertainty-aware blind image quality assessment in the laboratory and wild.IEEE Transactions on Image Pro- cessing, 30:3474–3486, 2021

    Weixia Zhang, Kede Ma, Guangtao Zhai, and Xiaokang Yang. Uncertainty-aware blind image quality assessment in the laboratory and wild.IEEE Transactions on Image Pro- cessing, 30:3474–3486, 2021. 6

  48. [48]

    Blind Image Quality Assessment via Vision- Language Correspondence: A Multitask Learning Perspec- tive

    Weixia Zhang, Guangtao Zhai, Ying Wei, Xiaokang Yang, and Kede Ma. Blind Image Quality Assessment via Vision- Language Correspondence: A Multitask Learning Perspec- tive. InIEEE Conference on Computer Vision and Pattern Recognition, pages 14071–14081, 2023. 6

  49. [49]

    Blind Super-resolution Quality Assessment based on a Resolution-adaptive Vision-language Model

    Weixia Zhang, Junlin Chen, Wei Sun, and Zhihua Wang. Blind Super-resolution Quality Assessment based on a Resolution-adaptive Vision-language Model. InProceedings of the IEEE/CVF Conference on Computer Vision (ICCV) Workshops, 2025. 6

  50. [50]

    Image Super-Resolution Using Very Deep Residual Channel Attention Networks

    Yulun Zhang, Kunpeng Li, Kai Li, Lichen Wang, Bineng Zhong, and Yun Fu. Image Super-Resolution Using Very Deep Residual Channel Attention Networks. InProceedings of the European Conference on Computer Vision, pages 286– 301, 2018. 2

  51. [51]

    Learning-Based Quality Assessment for Image Super-Resolution.IEEE Transactions on Multimedia, 24: 3570–3581, 2022

    Tiesong Zhao, Yuting Lin, Yiwen Xu, Weiling Chen, and Zhou Wang. Learning-Based Quality Assessment for Image Super-Resolution.IEEE Transactions on Multimedia, 24: 3570–3581, 2022. 1, 2

  52. [52]

    Visual Quality Assessment for Super-Resolved Images: Database and Method.IEEE Transactions on Image Pro- cessing, 28(7):3528–3541, 2019

    Fei Zhou, Rongguo Yao, Bozhi Liu, and Guoping Qiu. Visual Quality Assessment for Super-Resolved Images: Database and Method.IEEE Transactions on Image Pro- cessing, 28(7):3528–3541, 2019. 1

  53. [53]

    Quality Assessment of Image Super-Resolution: Balancing Deterministic and Statistical Fidelity

    Wei Zhou and Zhou Wang. Quality Assessment of Image Super-Resolution: Balancing Deterministic and Statistical Fidelity. InProceedings of the 30th ACM International Con- ference on Multimedia, pages 934–942, 2022. 1

  54. [54]

    Blind quality assessment for image superreso- lution using deep two-stream convolutional networks.Infor- mation Sciences, 528:205–218, 2020

    Wei Zhou, Qiuping Jiang, Yuwang Wang, Zhibo Chen, and Weiping Li. Blind quality assessment for image superreso- lution using deep two-stream convolutional networks.Infor- mation Sciences, 528:205–218, 2020. 1

  55. [55]

    Image Super- Resolution Quality Assessment: Structural Fidelity Versus Statistical Naturalness

    Wei Zhou, Zhou Wang, and Zhibo Chen. Image Super- Resolution Quality Assessment: Structural Fidelity Versus Statistical Naturalness. In2021 13th International confer- ence on quality of multimedia experience (QoMEX), pages 61–64. IEEE, 2021. 1

  56. [56]

    Perceptual Vi- sual Quality Assessment: Principles, Methods, and Future Directions.arXiv preprint arXiv:2503.00625, 2025

    Wei Zhou, Hadi Amirpour, Christian Timmerer, Guangtao Zhai, Patrick Le Callet, and Alan C Bovik. Perceptual Vi- sual Quality Assessment: Principles, Methods, and Future Directions.arXiv preprint arXiv:2503.00625, 2025. 1

  57. [57]

    VQualA 2025 Challenge on Visual Quality Comparison for Large Multimodal Mod- els: Methods and Results

    Hanwei Zhu, Haoning Wu, Zicheng Zhang, Lingyu Zhu, Yixuan Li, Peilin Chen, Shiqi Wang, Chris Wei Zhou, Lin- han Cao, Wei Sun, Xiangyang Zhu, Weixia Zhang, Yucheng Zhu, Jing Liu, Dandan Zhu, Guantao Zhai, Xiongkuo Min, Zhichao Zhang, Xinyue Li, Shubo Xu, Anh Dao, Yifan Li, Hongyuan Yu, Jiaojiao Yi, Yiding Tian, Yupeng Wu, Feiran Sun, Jiao Lijuan, and Song ...

This paper was first reviewed by deepseek-v4-flash on August 4, 2026.