Pith. sign in

REVIEW 4 major objections 5 minor 67 references

Efficient Face Image Quality Assessment via Self-training and Knowledge Distillation

T0 review · 4 major / 5 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read A distilled student model with 1.18 million parameters matches a Swin-B teacher on face image quality assessment (SRCC 0.9855 vs 0.9863) while using about one hundredth of the computation.

desk verdict Useful distillation recipe with a real challenge win, but the comparability claim rests on an unspecified eval split that doesn't square with the official test score. read the letter →

arxiv 2507.15709 v2 pith:US3OW5YC submitted 2025-07-21 cs.CV

Wei Sun , Weixia Zhang , Linhan Cao , Jun Jia , Xiangyang Zhu , Dandan Zhu , Xiongkuo Min , Guangtao Zhai This is my paper · ORCID
classification cs.CV
keywords faceimagequalityassessmentknowledgedistillationself-trainingpseudo-labelingmodelcompressionSwinTransformerEdgeNeXtVQualAchallenge
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 aims to show that a lightweight face image quality assessment (FIQA) model can match a much heavier one when the teacher is improved by self-training on its own pseudo-labels before distillation. The teacher, a Swin-B, is trained on 27,686 labeled face images, then used to pseudo-label 400,000 unlabeled Internet face images; those pseudo-labels both refine the teacher and supervise a compact student. The resulting EdgeNeXt-based student has 1.18M parameters and 0.4949 GFLOPs, and reaches SRCC 0.9855 / PLCC 0.9914 on the ICCV 2025 VQualA FIQA test set, within 0.001 of the teacher on both metrics. The method took first place in that challenge. This matters because accurate FIQA is currently too computationally heavy for many mobile and real-time applications.

What carries the argument

The engine is an iterative pseudo-labeling and distillation loop. A teacher trained on labeled data (Eq. 1) produces pseudo-labels for unlabeled images (Eq. 2); those pseudo-labels are added to the labeled set and used to fine-tune the teacher (Eq. 3), yielding an enhanced teacher that pseudo-labels a fresh set of unlabeled images; the student is then trained with a loss that sums labeled supervision, initial-teacher pseudo-labels, and enhanced-teacher pseudo-labels (Eq. 4). The regression loss combines MSE with a Pearson correlation (PLCC) term (Eqs. 5–7), which keeps predictions both close to the labels and rank-consistent with human ratings.

What would settle it

An experiment that would settle this: train the identical student on the labeled data alone, with no pseudo-labels, and compare its SRCC/PLCC on a fresh FIQA test set from a different source (e.g., synthetic distortions or a different demographic mix). If the gap between the distilled student and the teacher shrinks or reverses on that set, the pseudo-labeling gain is distribution-specific rather than general.

Watch

Extended reading notes

Core claim

The central claim is that a two-stage pseudo-labeling loop makes knowledge distillation work far better for FIQA: the teacher is first trained on labeled data, then used to pseudo-label a pool of unlabeled faces; those pseudo-labeled samples are fed back into the teacher to produce an enhanced teacher, which pseudo-labels a second pool; the student is trained on labeled data plus both pseudo-labeled pools. Ablations show each added source of pseudo-labels improves the student (SRCC 0.9807 with labeled only, 0.9845 with the first pseudo-label set, 0.9855 with both), and the enhanced teacher slightly outperforms the original (SRCC 0.9867 vs 0.9863). The student lands at 0.9855 SRCC / 0.9914 PLCC with 1.18M parameters and 0.4949 GFLOPs, a 99% parameter reduction and roughly 100x fewer FLOPs than the teacher, while staying within 0.001 of the teacher's accuracy.

Load-bearing premise

The load-bearing premise is that the pseudo-labels generated by the teacher on 400,000 unlabeled Internet face images are reliable enough and sufficiently aligned with the test distribution (Flickr-only images) that training on them improves both the enhanced teacher and the student.

Editorial extensions

If this is right

  • A student with 1.18M parameters and 0.4949 GFLOPs can replace a Swin-B teacher in deployment without a meaningful quality drop, enabling FIQA on mobile or real-time pipelines.
  • Adding pseudo-labels from the enhanced teacher (second stage) improves the student further than using only the initial teacher's pseudo-labels, so the extra pseudo-labeling round pays off.
  • The teacher's own accuracy rises slightly after self-training on its own pseudo-labels, so unlabeled Internet faces can supplement limited labeled FIQA data.
  • At 288x288 inference the model uses only 0.3313 GFLOPs, which is competitive with the other top challenge entries on both accuracy and cost.

Reading between the lines

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

  • Because the test set is exclusively Flickr images while unlabeled data is general Internet faces, the student's edge may shrink on other distributions (e.g., synthetic distortions or demographic subgroups); a cross-dataset test would reveal how much of the gain is distribution-specific.
  • The self-training loop assumes the teacher's pseudo-labels are unbiased; if they are systematically wrong on certain face types, the student inherits the bias. A simple probe would be to compare student vs teacher errors on low-quality or under-represented faces.
  • The same two-stage pseudo-labeling recipe could transfer to other perceptual regression tasks (e.g., video quality, aesthetic scoring) where unlabeled data is cheap and the teacher is large.
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 / 5 minor

Summary. The paper proposes a two-stage framework for efficient face image quality assessment (FIQA). A Swin-B teacher is first trained on labeled challenge data, then used to pseudo-label 200k unlabeled Internet face images. These pseudo-labels are used both to fine-tune an enhanced teacher (self-training) and to distill a lightweight EdgeNeXt student; a second pseudo-labeled set is generated by the enhanced teacher and also used for student training. The loss is a weighted combination of MSE and a PLCC-based term. The authors report that the student (1.18M parameters, 0.4949 GFLOPs) achieves SRCC 0.9855 / PLCC 0.9914 versus the teacher's 0.9863 / 0.9924, and that the method ranked first in the ICCV 2025 VQualA FIQA Challenge.

Significance. If the empirical claims hold, the paper offers a practically useful result: a very lightweight FIQA model with accuracy close to a much larger teacher, backed by a first-place challenge result and released code. The method is clearly described and the ablations in Table 2 give some support for the contribution of the pseudo-labeling stages. The significance is currently conditional, however, because the central teacher–student comparison in Table 1 is not tied to a specified evaluation split, and the reported official challenge score for the same student architecture is substantially lower. The single-run protocol also makes the small SRCC/PLCC differences difficult to interpret.

major comments (4)
  1. [§4.1 and Table 1] Table 1 is the only direct evidence for the core claim that the student is comparable to the teacher, but the paper never states which evaluation split this table reports. Section 4.1 defines a validation set (1,000 images, CelebA+Flickr) and a test set (889 images, Flickr-only), yet Table 1 gives only one SRCC/PLCC pair per method. The ambiguity is consequential: Table 4 reports an official challenge score of 0.9664 (the mean of SRCC and PLCC) for the same student architecture, whereas Table 1 implies a mean of about 0.9884. Section 4.7 says the challenge submission used 288×288 input while Table 1's student uses 352×352, but Table 3 shows that resolution changes of this magnitude affect SRCC by only about 0.001, so the 0.022 gap cannot be explained by resolution alone. The authors must specify the evaluation set for Table 1, report teacher and student results on both the validation and official test splits, and report the teacher's official test score; without this, the 'comparable performance' claim is not auditable.
  2. [§4.3] The sentence 'For fairness, all methods were retrained on the three test datasets, except NIQE' is not interpretable as written and, taken literally, describes training on test data, which would invalidate the comparisons in Table 1. Please clarify whether the intended meaning is that all methods were trained on the challenge training set and evaluated on both validation and test splits, and state exactly which data splits were used for training each baseline. This is essential for reproducibility of the baseline numbers.
  3. [§4.5, Tables 1 and 2] All reported results come from a single training run, with no error bars, confidence intervals, or significance tests. The central conclusion of teacher–student comparability rests on differences as small as 0.0008 in SRCC (teacher 0.9863 vs student 0.9855) and the claimed self-training gain is 0.0004 (teacher 0.9863 vs teacher+ 0.9867). These differences may be within run-to-run variation. Please report at least three independent training runs (mean ± std) for Tables 1 and 2, or otherwise justify the stability of these small differences.
  4. [§3.1 and §4.1] The self-training premise is that pseudo-labels produced by the teacher on roughly 400,000 unlabeled Internet face images are reliable and sufficiently aligned with the Flickr-only test distribution. The paper does not report any confidence filtering, domain alignment, or analysis of the pseudo-label distribution. Because both the enhanced teacher and the student are trained on these pseudo-labels, a systematic bias would be inherited by the student. Since the teacher's official test score is not reported, the paper currently cannot demonstrate that the self-training step improves generalization on the actual test distribution. I would like to see at least a distributional/agreement analysis of the pseudo-labels, or an evaluation of teacher vs teacher+ on the official test split, to support the claim that the self-training step helps.
minor comments (5)
  1. [Eq. (3)] The notation in Eq. (3) is malformed: the loss expression has unbalanced parentheses, and the symbol θt+ is not defined. It should probably be θ_t^+ (or similar), and the objective should be written with explicit parentheses around each loss term.
  2. [Eq. (4)] The third term in Eq. (4) uses ¯y_i without a definition. Please define ¯y_i as the pseudo-label generated by the enhanced teacher for the second unlabeled set, and ensure the notation matches Section 3.1.
  3. [Figure 1 caption] There is a typo in the caption: 'The the enhanced teacher model' should be 'The enhanced teacher model'.
  4. [§4.5] The sentence 'The lightweight baseline model, MobileNetV2, achieving moderate performance' is grammatically incomplete; it should read 'achieves moderate performance'.
  5. [Table 4] Table 4 reports only the overall challenge score, which is the average of SRCC and PLCC. If the individual SRCC and PLCC values for the challenge submission are available, they should be reported so that readers can compare the official test behavior with Table 1.

Circularity Check

0 steps flagged · score 1.0 of 10

No significant circularity: the student/teacher comparison is anchored to human MOS on challenge data, independent of the pseudo-label loop.

full rationale

The derivation chain is: the teacher is trained on human MOS labels (Eq. 1), generates pseudo-labels on unlabeled images (Eq. 2), is refined on labels plus pseudo-labels (Eq. 3), and the student is trained on labels plus pseudo-labels from both teacher stages (Eq. 4). The claimed result, namely that the student performs comparably to the teacher, is evaluated with SRCC/PLCC against human MOS on the challenge validation/test sets, which are external to the pseudo-labeled unlabeled corpora. The self-training loop uses the teacher's own outputs as auxiliary training targets, but the evaluation does not measure agreement with those pseudo-labels; it measures correlation with human ratings, so the teacher-student comparison is not forced by construction. The self-citations, such as the E-VQA inspiration in the Related Work section, are contextual and do not supply a load-bearing premise, and the VQualA first-place claim is an external ranking rather than a derived result. The reporting ambiguities noted by the skeptic, such as the unspecified evaluation split for Table 1 and the gap between Table 1 and the official challenge score, are correctness and reproducibility concerns rather than circularity.

Assumptions & free parameters 1 free parameters · 4 assumptions · 0 invented entities

No invented entities. One hand-chosen loss weight (lambda). The framework's performance rests on four domain assumptions: label reliability, pseudo-label reliability, unlabeled/test distribution alignment, and validation split reliability.

free parameters (1)
  • lambda (loss weight) = 1.0
    Weight balancing MSE and PLCC in Eq. (7); set by hand to 1, no tuning reported in the paper.
assumptions (4)
  • domain assumption Human MOS labels in the challenge training set reliably represent perceptual face image quality.
    The teacher is trained on these labels; if they are noisy or biased, the entire framework inherits the bias. Invoked in Eq. (1), Section 3.1.
  • domain assumption Pseudo-labels from the teacher on unlabeled images are a good enough proxy for human quality scores to improve both the teacher and the student.
    Core self-training premise; if pseudo-labels are systematically wrong on out-of-distribution images, the student inherits the error. Invoked in Eq. (2) and Eq. (4).
  • domain assumption The 400,000 unlabeled Internet face images share the distribution of the Flickr-only test set.
    Useful pseudo-labeling depends on distributional overlap; the data source is not detailed. Invoked in Section 4.1.
  • domain assumption The 80/20 train/private-validation split provides a reliable model selection signal.
    Best teachers are selected on this split; plausible with 27,686 labeled images but not demonstrated. Invoked in Section 4.1.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Efficient Face Image Quality Assessment via Self-training and Knowledge Distillation." pith.science (2026). https://pith.science/paper/US3OW5YC

@misc{pith2026250715709,
  author       = {Pith},
  title        = {Pith review of: Efficient Face Image Quality Assessment via Self-training and Knowledge Distillation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/US3OW5YC}},
  note         = {Machine review of arXiv:2507.15709}
}
read the original abstract

Face image quality assessment (FIQA) is essential for various face-related applications. Although FIQA has been extensively studied and achieved significant progress, the computational complexity of FIQA algorithms remains a key concern for ensuring scalability and practical deployment in real-world systems. In this paper, we aim to develop a computationally efficient FIQA method that can be easily deployed in real-world applications. Specifically, our method consists of two stages: training a powerful teacher model and distilling a lightweight student model from it. To build a strong teacher model, we adopt a self-training strategy to improve its capacity. We first train the teacher model using labeled face images, then use it to generate pseudo-labels for a set of unlabeled images. These pseudo-labeled samples are used in two ways: (1) to distill knowledge into the student model, and (2) to combine with the original labeled images to further enhance the teacher model through self-training. The enhanced teacher model is used to further pseudo-label another set of unlabeled images for distilling the student models. The student model is trained using a combination of labeled images, pseudo-labeled images from the original teacher model, and pseudo-labeled images from the enhanced teacher model. Experimental results demonstrate that our student model achieves comparable performance to the teacher model with an extremely low computational overhead. Moreover, our method achieved first place in the ICCV 2025 VQualA FIQA Challenge. The code is available at https://github.com/sunwei925/Efficient-FIQA.git.

Figures

Figures reproduced from arXiv: 2507.15709 by the authors.

Figure 1
Figure 1. The overall architecture of our proposed FIQA frame [PITH_FULL_IMAGE:figures/full_fig_p004_1.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

67 extracted references · 46 canonical work pages

  1. [1]

    Arniqa: Learning distortion mani- fold for image quality assessment

    Lorenzo Agnolucci, Leonardo Galteri, Marco Bertini, and Alberto Del Bimbo. Arniqa: Learning distortion mani- fold for image quality assessment. In Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision, pages 189–198, 2024. 2

    work page 2024

  2. [2]

    A fast approach for no- reference image sharpness assessment based on maximum local variation

    Khosro Bahrami and Alex C Kot. A fast approach for no- reference image sharpness assessment based on maximum local variation. IEEE Signal Processing Letters, 21(6):751– 755, 2014. 2

    work page 2014

  3. [3]

    Deep neural net- works for no-reference and full-reference image quality as- sessment

    Sebastian Bosse, Dominique Maniry, Klaus-Robert M ¨uller, Thomas Wiegand, and Wojciech Samek. Deep neural net- works for no-reference and full-reference image quality as- sessment. IEEE Transactions on Image Processing , 27(1): 206–219, 2017. 2

    work page 2017

  4. [4]

    Breaking annotation barriers: Generalized video quality assessment via ranking-based self- supervision

    Linhan Cao, Wei Sun, Kaiwei Zhang, Yicong Peng, Guang- tao Zhai, and Xiongkuo Min. Breaking annotation barriers: Generalized video quality assessment via ranking-based self- supervision. arXiv preprint arXiv:2505.03631, 2025. 2

    arXiv 2025

  5. [5]

    Attention-guided neural networks for full-reference and no- reference audio-visual quality assessment

    Yuqin Cao, Xiongkuo Min, Wei Sun, and Guangtao Zhai. Attention-guided neural networks for full-reference and no- reference audio-visual quality assessment. IEEE Transac- tions on Image Processing, 32:1882–1896, 2023. 2

    work page 2023

  6. [6]

    Agav-rater: adapting large multimodal model for ai-generated audio-visual quality assessment

    Yuqin Cao, Xiongkuo Min, Yixuan Gao, Wei Sun, and Guangtao Zhai. Agav-rater: adapting large multimodal model for ai-generated audio-visual quality assessment. arXiv preprint arXiv:2501.18314, 2025. 2

    arXiv 2025

  7. [7]

    An image quality assessment dataset for portraits

    Nicolas Chahine, Stefania Calarasanu, Davide Garcia- Civiero, Theo Cayla, Sira Ferradans, and Jean Ponce. An image quality assessment dataset for portraits. In Proceed- ings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 9968–9978, 2023. 3

    work page 2023

  8. [8]

    Topiq: A top-down approach from semantics to distortions for image quality assessment

    Chaofeng Chen, Jiadi Mo, Jingwen Hou, Haoning Wu, Liang Liao, Wenxiu Sun, Qiong Yan, and Weisi Lin. Topiq: A top-down approach from semantics to distortions for image quality assessment. IEEE Transactions on Image Processing, 33:2404–2418, 2024. 2

    work page 2024

Show all 67 references
  1. [9]

    Dsl-fiqa: As- sessing facial image quality via dual-set degradation learn- ing and landmark-guided transformer

    Wei-Ting Chen, Gurunandan Krishnan, Qiang Gao, Sy- Yen Kuo, Sizhou Ma, and Jian Wang. Dsl-fiqa: As- sessing facial image quality via dual-set degradation learn- ing and landmark-guided transformer. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recogn...

    2024

  2. [10]

    Vquala 2025 chal- lenge on genai-bench aigc video quality assessment: Meth- ods 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: Meth- ods and results. In Proceedings of the IEEE/CVF Conference on Computer Vision (ICCV) Workshops, pages 1–11, 2025. 2

    2025

  3. [11]

    A no-reference objective image sharpness metric based on the notion of just noticeable blur (jnb)

    Rony Ferzli and Lina J Karam. A no-reference objective image sharpness metric based on the notion of just noticeable blur (jnb). IEEE Transactions on Image Processing, 18(4): 717–728, 2009. 2

    2009

  4. [12]

    Born again neural net- works

    Tommaso Furlanello, Zachary Lipton, Michael Tschannen, Laurent Itti, and Anima Anandkumar. Born again neural net- works. In International Conference on Machine Learning , pages 1607–1616, 2018. 3

    2018

  5. [13]

    Lmm-vqa: Advancing video quality assessment with large multimodal models.IEEE Transactions on Circuits and Systems for Video Technology, 2025

    Qihang Ge, Wei Sun, Yu Zhang, Yunhao Li, Zhongpeng Ji, Fengyu Sun, Shangling Jui, Xiongkuo Min, and Guangtao Zhai. Lmm-vqa: Advancing video quality assessment with large multimodal models.IEEE Transactions on Circuits and Systems for Video Technology, 2025. 2

    2025

  6. [14]

    Faceqnet: Quality assessment for face recognition based on deep learning

    Javier Hernandez-Ortega, Javier Galbally, Julian Fierrez, Rudolf Haraksim, and Laurent Beslay. Faceqnet: Quality assessment for face recognition based on deep learning. In IEEE International Conference on Biometrics , pages 1–8,

  7. [15]

    Distill- ing the knowledge in a neural network

    Geoffrey Hinton, Oriol Vinyals, and Jeff Dean. Distill- ing the knowledge in a neural network. arXiv preprint arXiv:1503.02531, 2015. 3

    2015 arXiv

  8. [16]

    Vquala 2025 doc- ument image quality assessment challenge

    Fan Huang, Xiongkuo Min, Zhichao Ma, Xiaohong Liu, Chris Wei Zhou, Guangtao Zhai, and et al. Vquala 2025 doc- ument image quality assessment challenge. In Proceedings of the IEEE/CVF Conference on Computer Vision (ICCV) Workshops, pages 1–8, 2025. 2

    2025

  9. [17]

    Convolu- tional neural networks for no-reference image quality assess- ment

    Le Kang, Peng Ye, Yi Li, and David Doermann. Convolu- tional neural networks for no-reference image quality assess- ment. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pages 1733–1740, 2014. 2

    2014

  10. [18]

    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. 2, 5

    2021

  11. [19]

    Vquala 2025 challenge on engagement 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 ...

    2025

  12. [20]

    Vquala 2025 challenge on image super-resolution generated content qual- ity 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 content qual- ity assessment: Methods and results. In Proceed...

    2025

  13. [21]

    Aghi-qa: A subjective-aligned dataset and metric for ai-generated human images

    Yunhao Li, Sijing Wu, Wei Sun, Zhichao Zhang, Yucheng Zhu, Zicheng Zhang, Huiyu Duan, Xiongkuo Min, and Guangtao Zhai. Aghi-qa: A subjective-aligned dataset and metric for ai-generated human images. arXiv preprint arXiv:2504.21308, 2025. 2

    2025 arXiv

  14. [22]

    Assessing face image quality: A large-scale database and a transformer method

    Tie Liu, Shengxi Li, Mai Xu, Li Yang, and Xiaofei Wang. Assessing face image quality: A large-scale database and a transformer method. IEEE Transactions on Pattern Analysis and Machine Intelligence, 46(5):3981–4000, 2024. 1, 3

    2024

  15. [23]

    Swin transformer: Hierarchical vision transformer using shifted windows

    Ze Liu, Yutong Lin, Yue Cao, Han Hu, Yixuan Wei, Zheng Zhang, Stephen Lin, and Baining Guo. Swin transformer: Hierarchical vision transformer using shifted windows. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 10012–10022, 2021. 5

    2021

  16. [24]

    Bh-vqa: blind high frame rate video quality assessment

    Wei Lu, Wei Sun, Zicheng Zhang, Danyang Tu, Xiongkuo Min, and Guangtao Zhai. Bh-vqa: blind high frame rate video quality assessment. In IEEE International Conference on Multimedia and Expo, pages 2501–2506, 2023. 2

    2023

  17. [25]

    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 ...

    2025

  18. [26]

    Edgenext: efficiently amalgamated cnn-transformer architecture for mobile vision applications

    Muhammad Maaz, Abdelrahman Shaker, Hisham Cholakkal, Salman Khan, Syed Waqas Zamir, Rao Muham- mad Anwer, and Fahad Shahbaz Khan. Edgenext: efficiently amalgamated cnn-transformer architecture for mobile vision applications. In European Conference on Computer Vision, pages 3–2...

    2022

  19. [27]

    Image quality assessment us- ing contrastive learning

    Pavan C Madhusudana, Neil Birkbeck, Yilin Wang, Balu Adsumilli, and Alan C Bovik. Image quality assessment us- ing contrastive learning. IEEE Transactions on Image Pro- cessing, 31:4149–4161, 2022. 2

    2022

  20. [28]

    No-reference image quality assessment in the spa- tial domain

    Anish Mittal, Anush Krishna Moorthy, and Alan Conrad Bovik. No-reference image quality assessment in the spa- tial domain. IEEE Transactions on Image Processing , 21 (12):4695–4708, 2012. 2

    2012

  21. [29]

    completely blind

    Anish Mittal, Rajiv Soundararajan, and Alan C Bovik. Mak- ing a “completely blind” image quality analyzer. IEEE Sig- nal Processing Letters, 20(3):209–212, 2012. 2, 5

    2012

  22. [30]

    Blind im- age quality assessment: From natural scene statistics to per- ceptual quality

    Anush Krishna Moorthy and Alan Conrad Bovik. Blind im- age quality assessment: From natural scene statistics to per- ceptual quality. IEEE Transactions on Image Processing, 20 (12):3350–3364, 2011. 2, 5

    2011

  23. [31]

    A no-reference im- age blur metric based on the cumulative probability of blur detection (cpbd)

    Niranjan D Narvekar and Lina J Karam. A no-reference im- age blur metric based on the cumulative probability of blur detection (cpbd). IEEE Transactions on Image Processing , 20(9):2678–2683, 2011. 2

    2011

  24. [32]

    Sdd-fiqa: unsupervised face image quality assess- ment with similarity distribution distance

    Fu-Zhao Ou, Xingyu Chen, Ruixin Zhang, Yuge Huang, Shaoxin Li, Jilin Li, Yong Li, Liujuan Cao, and Yuan-Gen Wang. Sdd-fiqa: unsupervised face image quality assess- ment with similarity distribution distance. In Proceedings of the IEEE/CVF Conference on Computer Vision and Patt...

    2021

  25. [33]

    Statistics of natural images: Scaling in the woods

    Daniel L Ruderman and William Bialek. Statistics of natural images: Scaling in the woods. Physical review letters, 73(6): 814, 1994. 2

    1994

  26. [34]

    Mobilenetv2: Inverted residuals and linear bottlenecks

    Mark Sandler, Andrew Howard, Menglong Zhu, Andrey Zh- moginov, and Liang-Chieh Chen. Mobilenetv2: Inverted residuals and linear bottlenecks. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition , pages 4510–4520, 2018. 5

    2018

  27. [35]

    Face image quality assessment: A literature survey

    Torsten Schlett, Christian Rathgeb, Olaf Henniger, Javier Galbally, Julian Fierrez, and Christoph Busch. Face image quality assessment: A literature survey. ACM Computing Surveys, 54(10s):1–49, 2022. 1

    2022

  28. [36]

    Gray level co-occurrence matrices: generalisation and some new features

    Bino Sebastian V , A Unnikrishnan, and Kannan Balakrish- nan. Gray level co-occurrence matrices: generalisation and some new features. arXiv preprint arXiv:1205.4831, 2012. 2

    2012 arXiv

  29. [37]

    Going the extra mile in face image quality assess- ment: A novel database and model

    Shaolin Su, Hanhe Lin, Vlad Hosu, Oliver Wiedemann, Jin- qiu Sun, Yu Zhu, Hantao Liu, Yanning Zhang, and Dietmar Saupe. Going the extra mile in face image quality assess- ment: A novel database and model. IEEE Transactions on Multimedia, 26:2671–2685, 2023. 1, 2

    2023

  30. [38]

    Mc360iqa: A multi-channel cnn for blind 360-degree image quality assessment

    Wei Sun, Xiongkuo Min, Guangtao Zhai, Ke Gu, Huiyu Duan, and Siwei Ma. Mc360iqa: A multi-channel cnn for blind 360-degree image quality assessment. IEEE Journal of Selected Topics in Signal Processing, 14(1):64–77, 2019. 2

    2019

  31. [39]

    Deep learning based full-reference and no-reference quality assessment models for compressed ugc videos

    Wei Sun, Tao Wang, Xiongkuo Min, Fuwang Yi, and Guangtao Zhai. Deep learning based full-reference and no-reference quality assessment models for compressed ugc videos. In IEEE International Conference on Multimedia & Expo Workshops, pages 1–6, 2021. 2

    2021

  32. [40]

    A deep learning based no-reference quality assessment model for ugc videos

    Wei Sun, Xiongkuo Min, Wei Lu, and Guangtao Zhai. A deep learning based no-reference quality assessment model for ugc videos. In Proceedings of the 30th ACM Interna- tional Conference on Multimedia, pages 856–865, 2022. 2

    2022

  33. [41]

    Blind quality assessment for in-the-wild images via hierarchical feature fusion and iterative mixed database training

    Wei Sun, Xiongkuo Min, Danyang Tu, Siwei Ma, and Guangtao Zhai. Blind quality assessment for in-the-wild images via hierarchical feature fusion and iterative mixed database training. IEEE Journal of Selected Topics in Sig- nal Processing, 2023. 2, 5

    2023

  34. [42]

    Enhancing blind video quality as- sessment with rich quality-aware features

    Wei Sun, Haoning Wu, Zicheng Zhang, Jun Jia, Zhichao Zhang, Linhan Cao, Qiubo Chen, Xiongkuo Min, Weisi Lin, and Guangtao Zhai. Enhancing blind video quality as- sessment with rich quality-aware features. arXiv preprint arXiv:2405.08745, 2024. 3

    2024

  35. [43]

    Assessing uhd image quality from aesthetics, distor- tions, and saliency

    Wei Sun, Weixia Zhang, Yuqin Cao, Linhan Cao, Jun Jia, Zijian Chen, Zicheng Zhang, Xiongkuo Min, and Guangtao Zhai. Assessing uhd image quality from aesthetics, distor- tions, and saliency. In European Conference on Computer Vision, pages 109–126, 2024. 2

    2024

  36. [44]

    Dual-branch network for portrait image quality assessment

    Wei Sun, Weixia Zhang, Yanwei Jiang, Haoning Wu, Zicheng Zhang, Jun Jia, Yingjie Zhou, Zhongpeng Ji, Xiongkuo Min, Weisi Lin, et al. Dual-branch network for portrait image quality assessment. arXiv preprint arXiv:2405.08555, 2024. 3

    2024 arXiv

  37. [45]

    Compressedvqa-hdr: Generalized full-reference and no- reference quality assessment models for compressed high dynamic range videos

    Wei Sun, Linhan Cao, Kang Fu, Dandan Zhu, Jun Jia, Menghan Hu, Xiongkuo Min, and Guangtao Zhai. Compressedvqa-hdr: Generalized full-reference and no- reference quality assessment models for compressed high dynamic range videos. arXiv preprint arXiv:2506.22790 ,

    arXiv

  38. [46]

    An empirical study for efficient video quality assessment

    Wei Sun, Kang Fu, Linhan Cao, Dandan Zhu, Kaiwei Zhang, Yucheng Zhu, Zicheng Zhang, Menghan Hu, Xiongkuo Min, and Guangtao Zhai. An empirical study for efficient video quality assessment. In Proceedings of the Computer Vi- sion and Pattern Recognition Conference, pages 1403–1413,

  39. [47]

    Ser-fiq: Unsupervised esti- mation of face image quality based on stochastic embedding robustness

    Philipp Terhorst, Jan Niklas Kolf, Naser Damer, Florian Kirchbuchner, and Arjan Kuijper. Ser-fiq: Unsupervised esti- mation of face image quality based on stochastic embedding robustness. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pa...

    2020

  40. [48]

    S3: A spectral and spatial sharpness measure

    Cuong T Vu and Damon M Chandler. S3: A spectral and spatial sharpness measure. In International Conference on Advances in Multimedia, pages 37–43, 2009. 2

    2009

  41. [49]

    Ex- ploring clip for assessing the look and feel of images

    Jianyi Wang, Kelvin CK Chan, and Chen Change Loy. Ex- ploring clip for assessing the look and feel of images. InPro- ceedings of the AAAI Conference on Artificial Intelligence , pages 2555–2563, 2023. 2

    2023

  42. [50]

    Large multi-modality model assisted ai-generated image quality as- sessment

    Puyi Wang, Wei Sun, Zicheng Zhang, Jun Jia, Yanwei Jiang, Zhichao Zhang, Xiongkuo Min, and Guangtao Zhai. Large multi-modality model assisted ai-generated image quality as- sessment. In Proceedings of the 32nd ACM International Conference on Multimedia, pages 7803–7812, 2024. 2

    2024

  43. [51]

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

    Haoning Wu, Zicheng Zhang, Weixia Zhang, Chaofeng Chen, Liang Liao, Chunyi Li, Yixuan Gao, Annan Wang, Erli Zhang, Wenxiu Sun, et al. Q-align: Teaching lmms for visual scoring via discrete text-defined levels. In Inter- national Conference on Machine Learning , pages 54015– 54...

    2024

  44. [52]

    Fvq: A large-scale dataset and a lmm-based method for face video quality assessment

    Sijing Wu, Yunhao Li, Ziwen Xu, Yixuan Gao, Huiyu Duan, Wei Sun, and Guangtao Zhai. Fvq: A large-scale dataset and a lmm-based method for face video quality assessment. arXiv preprint arXiv:2504.09255, 2025. 2

    2025

  45. [53]

    Self-training with noisy student improves imagenet clas- sification

    Qizhe Xie, Minh-Thang Luong, Eduard Hovy, and Quoc V Le. Self-training with noisy student improves imagenet clas- sification. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages 10687– 10698, 2020. 3

    2020

  46. [54]

    Billion-scale semi-supervised learning for image classification

    I Zeki Yalniz, Herv ´e J ´egou, Kan Chen, Manohar Paluri, and Dhruv Mahajan. Billion-scale semi-supervised learning for image classification. arXiv preprint arXiv:1905.00546 ,

    1905 arXiv

  47. [55]

    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 , pag...

    2022

  48. [56]

    Attention based network for no- reference ugc video quality assessment

    Fuwang Yi, Mianyi Chen, Wei Sun, Xiongkuo Min, Yuan Tian, and Guangtao Zhai. Attention based network for no- reference ugc video quality assessment. In IEEE Interna- tional Conference on Image Processing , pages 1414–1418,

  49. [57]

    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. 2

    2020

  50. [58]

    A psychovisual quality metric in free-energy principle

    Guangtao Zhai, Xiaolin Wu, Xiaokang Yang, Weisi Lin, and Wenjun Zhang. A psychovisual quality metric in free-energy principle. IEEE Transactions on Image Processing , 21(1): 41–52, 2011. 2

    2011

  51. [59]

    Perceptual quality assessment of low-light image enhance- ment

    Guangtao Zhai, Wei Sun, Xiongkuo Min, and Jiantao Zhou. Perceptual quality assessment of low-light image enhance- ment. ACM Transactions on Multimedia Computing, Com- munications, and Applications, 17(4):1–24, 2021. 2

    2021

  52. [60]

    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 Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 14071–14081, 2023. 2, 5

    2023

  53. [61]

    A no-reference evalu- ation metric for low-light image enhancement

    Zicheng Zhang, Wei Sun, Xiongkuo Min, Wenhan Zhu, Tao Wang, Wei Lu, and Guangtao Zhai. A no-reference evalu- ation metric for low-light image enhancement. In IEEE In- ternational Conference on Multimedia and Expo, pages 1–6,

  54. [62]

    A no-reference deep learning quality assessment method for super-resolution im- ages based on frequency maps

    Zicheng Zhang, Wei Sun, Xiongkuo Min, Wenhan Zhu, Tao Wang, Wei Lu, and Guangtao Zhai. A no-reference deep learning quality assessment method for super-resolution im- ages based on frequency maps. In IEEE International Sym- posium on Circuits and Systems, pages 3170–3174, 2022. 2

    2022

  55. [63]

    Md-vqa: Multi-dimensional quality assessment for ugc live videos

    Zicheng Zhang, Wei Wu, Wei Sun, Danyang Tu, Wei Lu, Xiongkuo Min, Ying Chen, and Guangtao Zhai. Md-vqa: Multi-dimensional quality assessment for ugc live videos. In Proceedings of the IEEE/CVF Conference on Computer Vi- sion and Pattern Recognition, pages 1746–1755, 2023. 2

    2023

  56. [64]

    Benchmarking multi-dimensional aigc video qual- ity assessment: A dataset and unified model

    Zhichao Zhang, Wei Sun, Xinyue Li, Jun Jia, Xiongkuo Min, Zicheng Zhang, Chunyi Li, Zijian Chen, Puyi Wang, Fengyu Sun, et al. Benchmarking multi-dimensional aigc video qual- ity assessment: A dataset and unified model. arXiv preprint arXiv:2407.21408, 2024

    2024 arXiv

  57. [65]

    Human-activity agv quality assess- ment: A benchmark dataset and an objective evaluation met- ric

    Zhichao Zhang, Wei Sun, Xinyue Li, Yunhao Li, Qihang Ge, Jun Jia, Zicheng Zhang, Zhongpeng Ji, Fengyu Sun, Shangling Jui, et al. Human-activity agv quality assess- ment: A benchmark dataset and an objective evaluation met- ric. arXiv preprint arXiv:2411.16619, 2024. 2

    2024 arXiv

  58. [66]

    Advancing zero-shot digital human quality assessment through text-prompted evaluation

    Zicheng Zhang, Wei Sun, Yingjie Zhou, Haoning Wu, Chunyi Li, Xiongkuo Min, Xiaohong Liu, Guangtao Zhai, and Weisi Lin. Advancing zero-shot digital human quality assessment through text-prompted evaluation. IEEE Trans- actions on Image Processing, 2025. 2

    2025

  59. [67]

    Vquala 2025 challenge on visual quality comparison for large multimodal models: 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, Hong...

    2025

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.