Pith. sign in

REVIEW 4 major objections 4 minor 76 references

HiRQA: Hierarchical Ranking and Quality Alignment for Opinion-Unaware Image Quality Assessment

T0 review · 4 major / 4 minor · reviewed 2026-08-05 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read HiRQA claims that a no-reference image-quality model trained only on synthetic distortions, without human opinion labels, can score authentic unseen degradations such as lens flare, haze, motion blur, and low light.

desk verdict The submitted full text is not the HiRQA paper, so none of its claims can be checked; this artifact needs to go back to the authors, not to reviewers. read the letter →

arxiv 2508.15130 v3 pith:CWX6DUPE submitted 2025-08-20 cs.CV

Vaishnav Ramesh , Haining Wang , Md Jahidul Islam This is my paper
classification cs.CV
keywords no-referenceimagequalityassessmentopinion-unawareself-supervisedlearninghierarchicalrankinglossembeddingdistancetext-promptcontrastivealignmentsynthetic-to-authenticgeneralizationreal-timeinference
verification ladder T0 review T1 audit T2 compute T3 formal

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 proposes HiRQA, a no-reference image quality assessment framework that avoids the two most expensive inputs in the field: human opinion scores and a pristine reference image. Its central claim is that a network trained only on synthetic distortions can still produce quality scores that transfer to authentic, unseen degradations such as lens flare, haze, motion blur, and low light. Learning is driven by three signals: a higher-order ranking loss that orders distorted versions of the same content by relative quality, an embedding distance loss that ties feature similarity to perceived difference, and a contrastive alignment loss that uses structured text prompts. A lightweight variant, HiRQA-S, is reported to score an image in 3.5 milliseconds, placing no-reference quality assessment inside the budget of real-time pipelines. If the claim holds, quality scoring stops depending on collecting large human-rated datasets for every new deployment domain.

What carries the argument

The load-bearing component is the higher-order ranking loss: a loss that supervises quality predictions through the relational ordering of quality across distortion pairs (and their higher-order combinations) of the same content, instead of through absolute score labels. Because it only needs to know which distorted version looks better, it extracts quality supervision without human opinion scores. Two supporting mechanisms complete the embedding: an embedding distance loss that enforces consistency between feature distances and perceptual differences, and a contrastive alignment loss driven by structured textual prompts, which organizes the representation semantically at training time. The

What would settle it

Collect authentic photographs degraded by a mechanism absent from the synthetic training set — for instance, real lens flare or real atmospheric haze captured in the field — have human raters assign mean opinion scores, and compute the rank correlation between HiRQA's predictions and those human scores. If the correlation drops to chance level or falls well below the paper's reported synthetic-benchmark numbers, the generalization claim fails. A cheaper second check: ablate the higher-order ranking loss and the text-prompt alignment, retrain, and see whether the claimed authentic-distortion pe

Watch

Extended reading notes

Core claim

On the paper's own terms, the discovery is that relational ordering alone can teach absolute perceived quality. Rather than regressing to human mean opinion scores, HiRQA learns from the relative quality order among distortions of the same scene, supervised by a higher-order ranking loss, and from feature distances that are forced to match perceptual differences via an embedding distance loss. A training-time contrastive loss aligns the resulting embedding with structured text prompts, so the representation is semantically organized without ever seeing a label. At inference the model reads a quality score from the input image alone, and the authors report that training exclusively on synthet

Load-bearing premise

The load-bearing premise is that synthetic distortions, and the relative order of their severity, capture enough of real-world image quality that a model which has only seen them can correctly score authentic, unseen degradations.

Editorial extensions

If this is right

  • Training on purely synthetic distortions can produce quality scores that transfer to authentic, unseen degradations, removing the need to collect human opinion labels for each new deployment domain.
  • Quality scoring needs only the degraded image at inference, so the method applies where no reference image exists, such as user-generated photographs and real-time video.
  • The lightweight HiRQA-S variant's 3.5 ms per-image runtime keeps no-reference quality assessment within the budget of real-time capture and processing pipelines.
  • Because the embedding is aligned to structured text prompts, the learned representation carries semantic structure, not just a scalar score, which supports interpretability of what the model thinks is wrong with an image.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • Editorial inference: the sharpest test of the paper's premise is the class of authentic distortions with no close synthetic analog — artifacts from AI-generated imagery, complex optical aberrations, or sensor noise under extreme conditions — since those probe exactly the transfer gap being claimed.
  • Editorial inference: the text-prompt alignment suggests an untested extension: querying the frozen model with fine-grained textual descriptions of specific degradations to obtain per-distortion quality estimates without retraining.
  • Editorial inference: the ranking-plus-distance recipe is not specific to image quality; the same self-supervised objective could transfer to other label-scarce perceptual judgments, such as the quality of image enhancement or of generated content.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 4 minor

Summary. The submission is presented under the title and abstract of 'HiRQA: Hierarchical Ranking and Quality Alignment for Opinion-Unaware Image Quality Assessment,' claiming a self-supervised, opinion-unaware NR-IQA framework trained only on synthetic distortions, with three losses, competitive synthetic/authentic benchmark performance, and a 3.5 ms lightweight variant. The body of the manuscript, however, is not the HiRQA paper: it is arXiv:2508.15126, a paper describing the aiXiv open-access platform. The body contains no HiRQA method, no equations, no training details, no benchmark tables, and no results. Consequently, the central claims are unverifiable from the submitted artifact. The appended aiXiv text includes its own limitations (Secs. 5-6) disclaiming the status of AI-generated outputs, but those statements apply to aiXiv, not HiRQA.

Significance. If substantiated, the HiRQA contribution would be genuinely useful: an opinion-unaware NR-IQA model that transfers from synthetic to authentic distortions with no reference image at inference, plus a real-time variant, would address dataset bias and label dependence. The abstract's framing of synthetic-to-authentic generalization is a testable and valuable claim. However, the submission provides no evidence for this claim. No machine-checked proofs, reproducible code, or quantitative tables for HiRQA are present; the only reproducibility pointer is an external GitHub link. The manuscript in its current form cannot support any significance assessment.

major comments (4)
  1. [Full Text (all)] The body of the submission is not the HiRQA paper. It reproduces arXiv:2508.15126 ('aiXiv: A Next-Generation Open Access Ecosystem for Scientific Discovery'), whose abstract, method, experiments, and limitations concern an AI review platform. None of the HiRQA contribution—higher-order ranking loss, embedding distance loss, contrastive alignment loss, synthetic-only training, or evaluations on lens flare/haze/motion blur/low-light—appears in the body. This is a load-bearing structural defect: the abstract's claims are unsupported and unreviewable.
  2. [Abstract] No quantitative results are provided. The abstract asserts 'competitive performance,' 'strong generalization ability,' and an inference time of 3.5 ms, but no tables, metric values (e.g., SROCC/PLCC), benchmark names, or error bars are reported anywhere in the submission. The claim of comprehensive evaluations on unseen distortions is therefore uncheckable.
  3. [Abstract] The three losses are named but not defined. Equations for the 'higher-order ranking loss,' 'embedding distance loss,' and 'contrastive alignment loss' are absent, as is the training objective's weighting. Without these, the method's novelty and internal consistency cannot be assessed.
  4. [Abstract] The central transfer claim is an empirical assertion without protocol. The abstract says the model is 'trained only on synthetic image distortions' and generalizes to authentic degradations, but gives no training-set composition, distortion generation details, evaluation datasets, or comparison baselines. This needs a full experiments section before the claim can be evaluated.
minor comments (4)
  1. [Abstract] Define 'opinion-unaware' (e.g., no subjective opinion scores used in training) to avoid ambiguity with opinion-aware NR-IQA methods.
  2. [Full Text, Appendix B] The appendix contains sample preprints (DualScale Diffusion, Grokking, DualDiff, GAN-enhanced Diffusion) unrelated to HiRQA. If this manuscript is resubmitted, remove extraneous content.
  3. [Abstract] The GitHub link should include a commit hash and license; the repository is not part of the submitted manuscript and cannot be independently verified from the provided material.
  4. [Full Text, Sections 5-6] The aiXiv paper's limitations state that AI-generated outputs are preliminary drafts and that validation is confined to simulated settings. These are not HiRQA limitations, but they underscore that the appended text cannot substitute for HiRQA's methods/results.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity detectable: HiRQA's abstract contains no equations or fitted-parameter predictions, and the supplied full text is a different paper, so no derivation chain can be walked.

full rationale

The only HiRQA content available is its abstract. The abstract contains no equations, no fitted parameters, and no citations; every load-bearing claim — synthetic-only training, generalization to lens flare/haze/motion blur/low-light, and the 3.5 ms HiRQA-S inference time — is an empirical assertion about held-out conditions, not a quantity defined by the training signal. The 'higher-order ranking loss' and 'embedding distance loss' are named but not specified, so no identity between input and output can be exhibited, and no fitted input is renamed as a prediction. The supplied full text is the aiXiv platform paper (arXiv:2508.15126), not the HiRQA manuscript; that mismatch is an evidentiary gap, not a circular reduction. The aiXiv text's own limitations about simulated validation and preliminary AI-generated outputs further support the verifiability concern, but they do not establish that HiRQA's claims are circular. Because no derivation chain, self-citation load, or fitted-input-as-prediction structure is present in the provided text, there is no ground for a circularity finding; score 0 is the honest non-finding.

Assumptions & free parameters 0 free parameters · 3 assumptions · 0 invented entities

The ledger cannot be fully populated from the abstract because the full text is a different paper. We list the three domain assumptions explicitly stated or implied by the abstract. Any free parameters, such as loss weights or prompt templates, are unknown.

assumptions (3)
  • domain assumption Synthetic distortions form a sufficient training distribution for authentic degradation quality estimation.
    The abstract states 'Trained only on synthetic image distortions, HiRQA generalizes to authentic degradations.' If this transfer does not hold, the central generalization claim fails.
  • domain assumption Structured textual prompts provide a meaningful quality-prior signal for contrastive alignment.
    The abstract says 'A training-time contrastive alignment loss, guided by structured textual prompts, further enhances the learned representation.' This presumes text prompts are semantically aligned with perceptual quality, which is not established in the abstract.
  • domain assumption Higher-order relational ordering across distortion pairs is valid quality supervision.
    The abstract says 'We introduce a novel higher-order ranking loss that supervises quality predictions through relational ordering across distortion pairs.' This assumes that relative ordering of distorted versions is a sufficient proxy for absolute quality.

how reviews work

0 comments
Cite this review

Pith. "Pith review of HiRQA: Hierarchical Ranking and Quality Alignment for Opinion-Unaware Image Quality Assessment." pith.science (2026). https://pith.science/paper/CWX6DUPE

@misc{pith2026250815130,
  author       = {Pith},
  title        = {Pith review of: HiRQA: Hierarchical Ranking and Quality Alignment for Opinion-Unaware Image Quality Assessment},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/CWX6DUPE}},
  note         = {Machine review of arXiv:2508.15130}
}
read the original abstract

Despite significant progress in no-reference image quality assessment (NR-IQA), dataset biases and reliance on subjective labels continue to hinder their generalization performance. We propose HiRQA (Hierarchical Ranking and Quality Alignment), a self-supervised, opinion-unaware framework that offers a hierarchical, quality-aware embedding through a combination of ranking and contrastive learning. Unlike prior approaches that depend on pristine references or auxiliary modalities at inference time, HiRQA predicts quality scores using only the input image. We introduce a novel higher-order ranking loss that supervises quality predictions through relational ordering across distortion pairs, along with an embedding distance loss that enforces consistency between feature distances and perceptual differences. A training-time contrastive alignment loss, guided by structured textual prompts, further enhances the learned representation. Trained only on synthetic image distortions, HiRQA generalizes to authentic degradations, as demonstrated through comprehensive evaluations on various unseen distortions such as lens flare, haze, motion blur, and low-light conditions. For real-time deployment, we introduce HiRQA-S, a lightweight variant with an inference time of only 3.5 ms per image. Extensive experiments across synthetic and authentic benchmarks validate HiRQA's competitive performance, strong generalization ability, and scalability. The HiRQA model and inference pipeline are available at: https://github.com/uf-robopi/HiRQA.

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

76 extracted references · 64 canonical work pages

  1. [1]

    write newline

    " write newline "" before.all 'output.state := FUNCTION fin.entry add.period write newline FUNCTION new.block output.state before.all = 'skip after.block 'output.state := if FUNCTION new.sentence output.state after.block = 'skip output.state before.all = 'skip after.sentence 'output.state := if if FUNCTION not #0 #1 if FUNCTION and 'skip pop #0 if FUNCTIO...

  2. [2]

    write newline

    " write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 global.max substring 't := if while FUNCTION format.date year duplicate empty "emp...

  3. [3]

    # VC #ƄF -

    11em plus .33em minus .07em 4000 4000 100 4000 4000 500 `\.=1000 = #1 \@IEEEnotcompsoconly \@IEEEcompsoconly #1 * [1] 0pt [0pt][0pt] #1 * [1] 0pt [0pt][0pt] #1 * \| ** #1 \@IEEEauthorblockNstyle \@IEEEcompsocnotconfonly \@IEEEauthorblockAstyle \@IEEEcompsocnotconfonly \@IEEEcompsocconfonly \@IEEEauthordefaulttextstyle \@IEEEcompsocnotconfonly \@IEEEauthor...

    work page arXiv 1999

  4. [4]

    Quality-aware image-text alignment for opinion-unaware image quality assessment

    Lorenzo Agnolucci, Leonardo Galteri, and Marco Bertini. Quality-aware image-text alignment for opinion-unaware image quality assessment. arXiv preprint arXiv:2403.11176 , 2024

    arXiv 2024

  5. [5]

    Perceptual quality improvement in videoconferencing using keyframes-based gan

    Lorenzo Agnolucci, Leonardo Galteri, Marco Bertini, and Alberto Del Bimbo. Perceptual quality improvement in videoconferencing using keyframes-based gan. IEEE Transactions on Multimedia , 26:339--352, 2024

    work page 2024

  6. [6]

    Arniqa: Learning Distortion Manifold for Image Quality Assessment

    Lorenzo Agnolucci, Leonardo Galteri, Marco Bertini, and Alberto Del Bimbo. Arniqa: Learning Distortion Manifold for Image Quality Assessment . In Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision , pages 189--198, 2024

    work page 2024

  7. [7]

    I-HAZE: a dehazing benchmark with real hazy and haze-free indoor images

    Codruta O. Ancuti, Cosmin Ancuti, Radu Timofte, and Christophe De Vleeschouwer. I-HAZE: A Dehazing Benchmark With Real Hazy and Haze-Free Indoor Images . In arXiv:1804.05091v1 , 2018

    work page Pith review arXiv 2018

  8. [8]

    No reference opinion unaware quality assessment of authentically distorted images

    Nithin C Babu, Vignesh Kannan, and Rajiv Soundararajan. No reference opinion unaware quality assessment of authentically distorted images. In 2023 IEEE/CVF Winter Conference on Applications of Computer Vision (WACV) , pages 2458--2467, 2023

    work page 2023

Show all 76 references
  1. [9]

    On the Use of Deep Learning for Blind Image Quality Assessment

    Simone Bianco, Luigi Celona, Paolo Napoletano, and Raimondo Schettini. On the Use of Deep Learning for Blind Image Quality Assessment . Signal, Image and Video Processing , 12:355--362, 2018

    2018

  2. [10]

    Deep Neural Networks for No-Reference and Full-Reference Image Quality Assessment

    Sebastian Bosse, Dominique Maniry, Klaus-Robert M \"u ller, Thomas Wiegand, and Wojciech Samek. Deep Neural Networks for No-Reference and Full-Reference Image Quality Assessment . IEEE Transactions on image processing , 27(1):206--219, 2017

    2017

  3. [11]

    Learning to rank using gradient descent

    Christopher Burges, Tal Shaked, Erin Renshaw, Ari Lazier, Matt Deeds, Nicole Hamilton, and Gregory Hullender. Learning to rank using gradient descent. In ICML 2005 - Proceedings of the 22nd International Conference on Machine Learning , pages 89--96, 01 2005

    2005

  4. [12]

    A simple framework for contrastive learning of visual representations, 2020

    Ting Chen, Simon Kornblith, Mohammad Norouzi, and Geoffrey Hinton. A simple framework for contrastive learning of visual representations, 2020

    2020

  5. [13]

    Improved baselines with momentum contrastive learning, 2020

    Xinlei Chen, Haoqi Fan, Ross Girshick, and Kaiming He. Improved baselines with momentum contrastive learning, 2020

    2020

  6. [14]

    Iqagpt: Image quality assessment with vision-language and chatgpt models, 2023

    Zhihao Chen, Bin Hu, Chuang Niu, Tao Chen, Yuxin Li, Hongming Shan, and Ge Wang. Iqagpt: Image quality assessment with vision-language and chatgpt models, 2023

    2023

  7. [15]

    Grounding-iqa: Multimodal language grounding model for image quality assessment, 2025

    Zheng Chen, Xun Zhang, Wenbo Li, Renjing Pei, Fenglong Song, Xiongkuo Min, Xiaohong Liu, Xin Yuan, Yong Guo, and Yulun Zhang. Grounding-iqa: Multimodal language grounding model for image quality assessment, 2025

    2025

  8. [16]

    Deep Retinex Decomposition for Low-Light Enhancement

    Chen Wei, Wenjing Wang, Wenhan Yang, Jiaying Liu . Deep Retinex Decomposition for Low-Light Enhancement . In British Machine Vision Conference , 2018

    2018

  9. [17]

    Flare7K: A Phenomenological Nighttime Flare Removal Dataset

    Yuekun Dai, Chongyi Li, Shangchen Zhou, Ruicheng Feng, and Chen Change Loy. Flare7K: A Phenomenological Nighttime Flare Removal Dataset . In Thirty-sixth Conference on Neural Information Processing Systems Datasets and Benchmarks Track , 2022

    2022

  10. [18]

    Perceptual quality assessment of smartphone photography

    Yuming Fang, Hanwei Zhu, Yan Zeng, Kede Ma, and Zhou Wang. Perceptual quality assessment of smartphone photography. In IEEE Conference on Computer Vision and Pattern Recognition , pages 3677--3686, 2020

    2020

  11. [19]

    Universal Blind Image Quality Assessment Metrics Via Natural Scene Statistics and Multiple Kernel Learning

    Xinbo Gao, Fei Gao, Dacheng Tao, and Xuelong Li. Universal Blind Image Quality Assessment Metrics Via Natural Scene Statistics and Multiple Kernel Learning . IEEE Transactions on neural networks and learning systems , 24(12), 2013

    2013

  12. [20]

    Massive Online Crowdsourced Study of Subjective and Objective Picture Quality

    Deepti Ghadiyaram and Alan C Bovik. Massive Online Crowdsourced Study of Subjective and Objective Picture Quality . IEEE Transactions on Image Processing , 25(1):372--387, 2015

    2015

  13. [21]

    Perceptual Quality Prediction on Authentically Distorted Images Using a Bag of Features Approach

    Deepti Ghadiyaram and Alan C Bovik. Perceptual Quality Prediction on Authentically Distorted Images Using a Bag of Features Approach . Journal of vision , 17(1):32--32, 2017

    2017

  14. [22]

    No-Reference Image Quality Assessment Via Transformers, Relative Ranking, and Self-Consistency

    S Alireza Golestaneh, Saba Dadsetan, and Kris M Kitani. No-Reference Image Quality Assessment Via Transformers, Relative Ranking, and Self-Consistency . In Proceedings of the IEEE/CVF winter conference on applications of computer vision , pages 1220--1230, 2022

    2022

  15. [23]

    Deep residual learning for image recognition, 2015

    Kaiming He, Xiangyu Zhang, Shaoqing Ren, and Jian Sun. Deep residual learning for image recognition, 2015

    2015

  16. [24]

    KonIQ-10k: An Ecologically Valid Database for Deep Learning of Blind Image Quality Assessment

    Vlad Hosu, Hanhe Lin, Tamas Sziranyi, and Dietmar Saupe. KonIQ-10k: An Ecologically Valid Database for Deep Learning of Blind Image Quality Assessment . IEEE Transactions on Image Processing , 29:4041--4056, 2020

    2020

  17. [25]

    Tongue Image Quality Assessment Based on a Deep Convolutional Neural Network

    Tao Jiang, Xiao-juan Hu, Xing-hua Yao, Li-ping Tu, Jing-bin Huang, Xu-xiang Ma, Ji Cui, Qing-feng Wu, and Jia-tuo Xu. Tongue Image Quality Assessment Based on a Deep Convolutional Neural Network . BMC Medical Informatics and Decision Making , 21(1):147, 2021

    2021

  18. [26]

    Convolutional Neural Networks for No-Reference Image Quality Assessment

    Le Kang, Peng Ye, Yi Li, and David Doermann. Convolutional Neural Networks for No-Reference Image Quality Assessment . In Proceedings of the IEEE conference on computer vision and pattern recognition , pages 1733--1740, 2014

    2014

  19. [27]

    Musiq: Multi-Scale Image Quality Transformer

    Junjie Ke, Qifei Wang, Yilin Wang, Peyman Milanfar, and Feng Yang. Musiq: Multi-Scale Image Quality Transformer . In Proceedings of the IEEE/CVF international conference on computer vision , pages 5148--5157, 2021

    2021

  20. [28]

    Most Apparent Distortion: Full-Reference Image Quality Assessment and The Role of Strategy

    Eric C Larson and Damon M Chandler. Most Apparent Distortion: Full-Reference Image Quality Assessment and The Role of Strategy . Journal of electronic imaging , 19(1):011006--011006, 2010

    2010

  21. [29]

    Dongxu Li, Junnan Li, Hung Le, Guangsen Wang, Silvio Savarese, and Steven C.H. Hoi. LAVIS : A one-stop library for language-vision intelligence. In Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 3: System Demonstrations) , pages...

    2023

  22. [30]

    Blip: Bootstrapping language-image pre-training for unified vision-language understanding and generation

    Junnan Li, Dongxu Li, Caiming Xiong, and Steven Hoi. Blip: Bootstrapping language-image pre-training for unified vision-language understanding and generation. In ICML , 2022

    2022

  23. [31]

    Boosting clip adaptation for image quality assessment via meta-prompt learning and gradient regularization, 2024

    Xudong Li, Zihao Huang, Runze Hu, Yan Zhang, Liujuan Cao, and Rongrong Ji. Boosting clip adaptation for image quality assessment via meta-prompt learning and gradient regularization, 2024

    2024

  24. [32]

    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 . In Proceedings of the IEEE/CVF international conference on computer vision , pages 1833--1844, 2021

    2021

  25. [33]

    KADID-10k: A Large-Scale Artificially Distorted IQA Database

    Hanhe Lin, Vlad Hosu, and Dietmar Saupe. KADID-10k: A Large-Scale Artificially Distorted IQA Database . In Eleventh International Conference on Quality of Multimedia Experience (QoMEX) , pages 1--3. IEEE, 2019

    2019

  26. [34]

    Deepfl-iqa: Weak supervision for deep iqa feature learning

    Hanhe Lin, Vlad Hosu, and Dietmar Saupe. Deepfl-iqa: Weak supervision for deep iqa feature learning. arXiv preprint arXiv:2001.08113 , 2020

    2001 arXiv

  27. [35]

    Hallucinated-IQA: No-Reference Image Quality Assessment Via Adversarial Learning

    Kwan-Yee Lin and Guanxiang Wang. Hallucinated-IQA: No-Reference Image Quality Assessment Via Adversarial Learning . In Proceedings of the IEEE conference on computer vision and pattern recognition , pages 732--741, 2018

    2018

  28. [36]

    dipIQ: Blind Image Quality Assessment by Learning-to-Rank Discriminable Image Pairs

    Kede Ma, Wentao Liu, Tongliang Liu, Zhou Wang, and Dacheng Tao. dipIQ: Blind Image Quality Assessment by Learning-to-Rank Discriminable Image Pairs . IEEE Transactions on Image Processing , 26(8):3951--3964, 2017

    2017

  29. [37]

    Image quality assessment using contrastive learning

    Pavan C Madhusudana, Neil Birkbeck, Yilin Wang, Balu Adsumilli, and Alan C Bovik. Image quality assessment using contrastive learning. arXiv:2110.13266 , 2021

    2021 arXiv

  30. [38]

    No-Reference Image Qality Assessment in the Spatial Domain

    Anish Mittal, Anush Krishna Moorthy, and Alan Conrad Bovik. No-Reference Image Qality Assessment in the Spatial Domain . IEEE Transactions on image processing , 21(12):4695--4708, 2012

    2012

  31. [39]

    completely blind

    Anish Mittal, Rajiv Soundararajan, and Alan C. Bovik. Making a “completely blind” image quality analyzer. IEEE Signal Processing Letters , 20(3):209--212, 2013

    2013

  32. [40]

    Blind Image Quality Assessment: From Natural Scene Statistics to Perceptual Quality

    Anush Krishna Moorthy and Alan Conrad Bovik. Blind Image Quality Assessment: From Natural Scene Statistics to Perceptual Quality . IEEE transactions on Image Processing , 20(12):3350--3364, 2011

    2011

  33. [41]

    Deep Multi-Scale Convolutional Neural Network for Dynamic Scene Deblurring

    Seungjun Nah, Tae Hyun Kim, and Kyoung Mu Lee. Deep Multi-Scale Convolutional Neural Network for Dynamic Scene Deblurring . In Proceedings of the IEEE conference on computer vision and pattern recognition , pages 3883--3891, 2017

    2017

  34. [42]

    Opinion-unaware blind image quality assessment using multi-scale deep feature statistics, 2024

    Zhangkai Ni, Yue Liu, Keyan Ding, Wenhan Yang, Hanli Wang, and Shiqi Wang. Opinion-unaware blind image quality assessment using multi-scale deep feature statistics, 2024

    2024

  35. [43]

    Image Database TID2013: Peculiarities, Results and Perspectives

    Nikolay Ponomarenko, Lina Jin, Oleg Ieremeiev, Vladimir Lukin, Karen Egiazarian, Jaakko Astola, Benoit Vozel, Kacem Chehdi, Marco Carli, Federica Battisti, et al. Image Database TID2013: Peculiarities, Results and Perspectives . Signal processing: Image communication , 30:57--77, 2015

    2015

  36. [44]

    Learning transferable visual models from natural language supervision, 2021

    Alec Radford, Jong Wook Kim, Chris Hallacy, Aditya Ramesh, Gabriel Goh, Sandhini Agarwal, Girish Sastry, Amanda Askell, Pamela Mishkin, Jack Clark, Gretchen Krueger, and Ilya Sutskever. Learning transferable visual models from natural language supervision, 2021

    2021

  37. [45]

    Dgiqa: Depth-guided feature attention and refinement for generalizable image quality assessment, 2025

    Vaishnav Ramesh, Junliang Liu, Haining Wang, and Md Jahidul Islam. Dgiqa: Depth-guided feature attention and refinement for generalizable image quality assessment, 2025

    2025

  38. [46]

    Blind Image Quality Assessment: A Natural Scene Statistics Approach in the DCT Domain

    Michele A Saad, Alan C Bovik, and Christophe Charrier. Blind Image Quality Assessment: A Natural Scene Statistics Approach in the DCT Domain . IEEE transactions on Image Processing , 21(8):3339--3352, 2012

    2012

  39. [47]

    Re-iqa: Unsupervised Learning for Image Quality Assessment In the Wild

    Avinab Saha, Sandeep Mishra, and Alan C Bovik. Re-iqa: Unsupervised Learning for Image Quality Assessment In the Wild . In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages 5846--5855, 2023

    2023

  40. [48]

    A Statistical Evaluation of Recent Full Reference Image Quality Assessment Algorithms

    Hamid R Sheikh, Muhammad F Sabir, and Alan C Bovik. A Statistical Evaluation of Recent Full Reference Image Quality Assessment Algorithms . IEEE Transactions on image processing , 15(11):3440--3451, 2006

    2006

  41. [49]

    Learning generalizable perceptual representations for data-efficient no-reference image quality assessment

    Suhas Srinath, Shankhanil Mitra, Shika Rao, and Rajiv Soundararajan. Learning generalizable perceptual representations for data-efficient no-reference image quality assessment. In Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision (WACV) , pages 2...

    2024

  42. [50]

    Blindly Assess Image Quality in the Wild Guided by a Self-Adaptive Hyper Network

    Shaolin Su, Qingsen Yan, Yu Zhu, Cheng Zhang, Xin Ge, Jinqiu Sun, and Yanning Zhang. Blindly Assess Image Quality in the Wild Guided by a Self-Adaptive Hyper Network . In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages 3667--3676, 2020

    2020

  43. [51]

    NIMA: Neural Image Assessment

    Hossein Talebi and Peyman Milanfar. NIMA: Neural Image Assessment . IEEE transactions on image processing , 27(8):3998--4011, 2018

    2018

  44. [52]

    Representation learning with contrastive predictive coding, 2019

    Aaron van den Oord, Yazhe Li, and Oriol Vinyals. Representation learning with contrastive predictive coding, 2019

    2019

  45. [53]

    Visualizing Data Using t-SNE

    Laurens Van der Maaten and Geoffrey Hinton. Visualizing Data Using t-SNE. Journal of machine learning research , 9(11), 2008

    2008

  46. [54]

    Exploring Clip for Assessing the Look and Feel of Images

    Jianyi Wang, Kelvin CK Chan, and Chen Change Loy. Exploring Clip for Assessing the Look and Feel of Images . In Proceedings of the AAAI Conference on Artificial Intelligence , volume 37, pages 2555--2563, 2023

    2023

  47. [55]

    A Survey on IQA

    Lanjiang Wang. A Survey on IQA . arXiv preprint arXiv:2109.00347 , 2021

    2021 arXiv

  48. [56]

    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 . In Proceedings of the IEEE/CVF international conference on computer vision , pages 1905--1914, 2021

    1905

  49. [57]

    Modern Image Quality Assessment

    Zhou Wang and Alan Conrad Bovik. Modern Image Quality Assessment . PhD thesis, Springer, 2006

    2006

  50. [58]

    Why Is Image Quality Assessment So Difficult? In IEEE International conference on acoustics, speech, and signal processing , volume 4, pages IV--3313

    Zhou Wang, Alan C Bovik, and Ligang Lu. Why Is Image Quality Assessment So Difficult? In IEEE International conference on acoustics, speech, and signal processing , volume 4, pages IV--3313. IEEE, 2002

    2002

  51. [59]

    Image Quality Assessment: From Error Visibility to Structural Similarity

    Zhou Wang, Alan C Bovik, Hamid R Sheikh, and Eero P Simoncelli. Image Quality Assessment: From Error Visibility to Structural Similarity . IEEE transactions on image processing , 13(4):600--612, 2004

    2004

  52. [60]

    Q-bench: A benchmark for general-purpose foundation models on low-level vision, 2024

    Haoning Wu, Zicheng Zhang, Erli Zhang, Chaofeng Chen, Liang Liao, Annan Wang, Chunyi Li, Wenxiu Sun, Qiong Yan, Guangtao Zhai, and Weisi Lin. Q-bench: A benchmark for general-purpose foundation models on low-level vision, 2024

    2024

  53. [61]

    Q-align: Teaching lmms for visual scoring via discrete text-defined levels

    Haoning Wu, Zicheng Zhang, Weixia Zhang, Chaofeng Chen, Chunyi Li, Liang Liao, Annan Wang, Erli Zhang, Wenxiu Sun, Qiong Yan, Xiongkuo Min, Guangtai Zhai, and Weisi Lin. Q-align: Teaching lmms for visual scoring via discrete text-defined levels. arXiv preprint arXiv:2312.17090...

    2023 arXiv

  54. [62]

    A Comprehensive Study of Multimodal Large Language Models for Image Quality Assessment

    Tianhe Wu, Kede Ma, Jie Liang, Yujiu Yang, and Lei Zhang. A Comprehensive Study of Multimodal Large Language Models for Image Quality Assessment . arXiv preprint arXiv:2403.10854 , 2024

    2024 arXiv

  55. [63]

    Blind Image Quality Assessment Based on High Order Statistics Aggregation

    Jingtao Xu, Peng Ye, Qiaohong Li, Haiqing Du, Yong Liu, and David Doermann. Blind Image Quality Assessment Based on High Order Statistics Aggregation . IEEE Transactions on Image Processing , 25(9):4444--4457, 2016

    2016

  56. [64]

    Maniqa: Multi-Dimension Attention Network for No-Reference Image Quality Assessment

    Sidi Yang, Tianhe Wu, Shuwei Shi, Shanshan Lao, Yuan Gong, Mingdeng Cao, Jiahao Wang, and Yujiu Yang. Maniqa: Multi-Dimension Attention Network for No-Reference Image Quality Assessment . In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pa...

    2022

  57. [65]

    No-Reference Image Quality Assessment Using Visual Codebooks

    Peng Ye and David Doermann. No-Reference Image Quality Assessment Using Visual Codebooks . IEEE Transactions on Image Processing , 21(7):3129--3138, 2012

    2012

  58. [66]

    Unsupervised Feature Learning Framework for No-Reference Image Quality Assessment

    Peng Ye, Jayant Kumar, Le Kang, and David Doermann. Unsupervised Feature Learning Framework for No-Reference Image Quality Assessment . In IEEE conference on computer vision and pattern recognition , pages 1098--1105. IEEE, 2012

    2012

  59. [67]

    From Patches to Pictures (PaQ-2-PiQ): Mapping the Perceptual Space of Picture Quality

    Zhenqiang Ying, Haoran Niu, Praful Gupta, Dhruv Mahajan, Deepti Ghadiyaram, and Alan Bovik. From Patches to Pictures (PaQ-2-PiQ): Mapping the Perceptual Space of Picture Quality . In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages 3575...

    2020

  60. [68]

    Transformer for Image Quality Assessment

    Junyong You and Jari Korhonen. Transformer for Image Quality Assessment . In IEEE international conference on image processing (ICIP) , pages 1389--1393. IEEE, 2021

    2021

  61. [69]

    Depicting beyond scores: Advancing image quality assessment through multi-modal language models, 2024

    Zhiyuan You, Zheyuan Li, Jinjin Gu, Zhenfei Yin, Tianfan Xue, and Chao Dong. Depicting beyond scores: Advancing image quality assessment through multi-modal language models, 2024

    2024

  62. [70]

    Perceptual Image Quality Assessment: A Survey

    Guangtao Zhai and Xiongkuo Min. Perceptual Image Quality Assessment: A Survey . Science China Information Sciences , 63:1--52, 2020

    2020

  63. [71]

    Lin Zhang, Lei Zhang, and Alan C. Bovik. A feature-enriched completely blind image quality evaluator. IEEE Transactions on Image Processing , 24(8):2579--2591, 2015

    2015

  64. [72]

    Blind Image Quality Assessment Using a Deep Bilinear Convolutional Neural Network

    Weixia Zhang, Kede Ma, Jia Yan, Dexiang Deng, and Zhou Wang. Blind Image Quality Assessment Using a Deep Bilinear Convolutional Neural Network . IEEE Transactions on Circuits and Systems for Video Technology , 30(1):36--47, 2018

    2018

  65. [73]

    Blind image quality assessment via vision-language correspondence: A multitask learning perspective

    Weixia Zhang, Guangtao Zhai, Ying Wei, Xiaokang Yang, and Kede Ma. Blind image quality assessment via vision-language correspondence: A multitask learning perspective. In IEEE Conference on Computer Vision and Pattern Recognition , pages 14071--14081, 2023

    2023

  66. [74]

    Quality-aware pretrained models for blind image quality assessment

    Kai Zhao, Kun Yuan, Ming Sun, Mading Li, and Xing Wen. Quality-aware pretrained models for blind image quality assessment. In CVPR , pages 22302--22313. IEEE , 2023

    2023

  67. [75]

    Vision language modeling of content, distortion and appearance for image quality assessment, 2024

    Fei Zhou, Tianhao Gu, Zhicong Huang, and Guoping Qiu. Vision language modeling of content, distortion and appearance for image quality assessment, 2024

    2024

  68. [76]

    MetaIQA: Deep Meta-Learning for No-Reference Image Quality Assessment

    Hancheng Zhu, Leida Li, Jinjian Wu, Weisheng Dong, and Guangming Shi. MetaIQA: Deep Meta-Learning for No-Reference Image Quality Assessment . In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages 14143--14152, 2020

    2020

Pith tools

Reviewed August 5, 2026 · model on record in the stance chip above.