Pith. sign in

REVIEW 4 major objections 5 minor 1 cited by

DiffVQA: Video Quality Assessment Using Diffusion Feature Extractor

T0 review · 4 major / 5 minor · reviewed 2026-08-15 · deepseek-v4-flash

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

Pith's one-line read No-reference video quality assessment improves when a diffusion model supplies the features, and the paper reports gains over CNN and ViT backbones on three UGC benchmarks.

desk verdict Strong empirical results on five VQA datasets, but the central claim about diffusion features rests on a confounded backbone comparison and a shortcut-prone adaptation objective. read the letter →

arxiv 2505.03261 v1 pith:35OQLLSB submitted 2025-05-06 cs.CV eess.IV

Wei-Ting Chen , Yu-Jiet Vong , Yi-Tsung Lee , Sy-Yen Kuo , Qiang Gao , Sizhuo Ma , Jian Wang This is my paper · ORCID
classification cs.CVeess.IV
keywords videoqualityassessmentno-referenceVQAdiffusionfeatureextractorStableMambastatespacemodeluser-generatedcontentcross-datasetgeneralization
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

This paper claims that a latent diffusion model adapted to reconstruct video frames can serve as a feature extractor for no-reference video quality assessment, and that the resulting features align with human quality judgments better than features from CNNs or vision transformers. On LIVE-VQC, KoNViD-1k, and YouTube-UGC, the proposed DiffVQA model reports Spearman correlations of 0.910, 0.936, and 0.923, improving on prior methods by as much as roughly 0.05. The same model also transfers across datasets without fine-tuning, which the authors take as evidence that diffusion priors trained on large image collections carry useful quality-relevant information. If the claim holds, adapting generative models for reconstruction becomes a viable route to perceptual feature extraction, not just synthesis.

What carries the argument

The central object is the adapted diffusion feature extractor: a pre-trained latent diffusion model whose denoising U-Net is augmented by a Controller, trained with the loss $L_{\text{Adap}}=\mathbb{E}_{\mathbf{z}_0,t,\epsilon}[\lVert \epsilon_\theta(\mathbf{z}_t,t,\mathbf{z}_0)-\epsilon\rVert^2]$. The load-bearing identity is the reconstruction objective: the paper assumes that if the model reconstructs each input frame closely, its internal features must encode the semantic and distortion information needed for quality scoring. The parallel temporal mechanism is the Temporal Coherence Augmentation Block, built on bidirectional Mamba state-space scans, which supplies long-range temporal coherence features that are merged with the diffusion features before score prediction.

What would settle it

Train the identical VQA head and feature-fusion pipeline on features obtained directly from the VAE encoder without the denoising U-Net, or retrain with the Controller given only $\mathbf{z}_t$ and not $\mathbf{z}_0$; if the reported correlation scores over ViT backbones persist, the diffusion prior is not the cause, and if they collapse, the reconstruction shortcut is doing the work.

Watch

Extended reading notes

Core claim

DiffVQA adapts Stable Diffusion by adding a trainable Controller that takes the noisy latent and the clean latent as input and is trained to denoise the noisy latent back to the original frame; after adaptation, the denoising network's multi-scale features at time step zero plus the reconstructed latent are treated as quality features. A resizing branch yields semantic features and a random-crop branch yields distortion features, and both are fused with temporal-difference features. A parallel bidirectional Mamba block, the Temporal Coherence Augmentation Block, adds long-range temporal coherence features, and the final video score is a learned weighted sum of per-frame quality scores. The paper reports that this setup outperforms CNN, ViT, CLIP, and VAE backbones inside the same architecture, both within datasets and in cross-dataset transfer.

Load-bearing premise

The whole method rests on the assumption that because the adapted diffusion model can reconstruct each input frame, its internal features must contain the semantic and distortion information needed for quality scoring; since the clean latent $\mathbf{z}_0$ is concatenated into the Controller's input, the model may succeed by copying $\mathbf{z}_0$ rather than by learning generalizable quality features.

Editorial extensions

If this is right

  • Diffusion-based feature extractors can replace CNN and ViT backbones in no-reference VQA and improve both same-dataset accuracy and cross-dataset transfer.
  • The parallel Mamba block offers a stronger temporal modeling alternative to 3D convolutions for long video sequences within the same pipeline.
  • Splitting features into semantic and distortion branches and learning per-frame weights makes the quality prediction combine spatial content and distortion cues explicitly.
  • Because the feature extractor transfers across datasets, small VQA datasets can be fine-tuned without sacrificing cross-dataset generalization.

Reading between the lines

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

  • A control the paper does not run: remove the clean latent $\mathbf{z}_0$ from the Controller's input and retrain; if performance drops, the gains come from the copy route rather than from diffusion reconstruction.
  • Comparing against a frozen, unadapted diffusion feature extractor would separate the contribution of large-scale pretraining from the contribution of the reconstruction adaptation.
  • The same adapt-then-extract recipe could transfer to other perceptual tasks with small datasets, but the copy-through-conditioning pitfall would need to be controlled first.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

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 introduces DiffVQA, a no-reference video quality assessment model that adapts a Stable Diffusion U-Net with a ControlNet-style controller to reconstruct input frames, then uses the adapted model as a frozen feature extractor. Features are obtained from a resized branch (semantics) and a random-crop branch (distortion), fused with a Mamba-based temporal coherence augmentation block, and combined via a weighted MLP to predict video quality scores. Experiments on KoNViD-1k, LIVE-VQC, YouTube-UGC, LSVQ, and KVQ report state-of-the-art intra-dataset SRCC/PLCC and strong cross-dataset generalization, with ablations comparing the diffusion extractor against CNN, ViT, CLIP, Vision Mamba, and VAE backbones.

Significance. If the central causal claim holds, the paper is significant: it demonstrates that a large pre-trained diffusion model can be repurposed as a feature extractor for perceptual quality, with gains of roughly 0.02-0.05 SRCC over prior methods on several benchmarks and notable cross-dataset robustness. The paper also provides a broad backbone comparison and a Mamba-based temporal module, and the experiments are extensive across five datasets. However, the significance hinges on the ablations being able to attribute the gains to the diffusion prior rather than to the extra task-specific adaptation that the diffusion extractor receives; as reported, the evidence is underdetermined.

major comments (4)
  1. [Section 4.2, Table 5] The comparison in Table 5 is confounded by unequal training budgets. The diffusion feature extractor receives 30,000 iterations of Controller adaptation on the target VQA datasets (Section 4.2), while the ResNet, ViT, CLIP, and Vision Mamba baselines are used only with pre-trained weights. The text states that 'All variants are trained under identical conditions,' but the diffusion variant has an additional task-specific adaptation stage that the other backbones do not. The LIVE-VQC SRCC gap between DiffVQA (0.910) and ViT-L/16 (0.857) could therefore reflect adaptation to the evaluation distribution rather than the diffusion prior. Please add controls that give comparable adaptation to the other backbones (for example, a reconstruction-based or quality-based fine-tuning stage with a similar iteration budget) and also report the performance of the diffusion extractor without the 30,000-iteration adaptation.
  2. [Section 3.2, Eq. (10)] The principal mechanism is asserted rather than derived. In Eq. (10), the Controller receives the concatenation of the noisy latent z_t and the clean latent z_0 as conditioning, so the reconstruction task can be solved almost by copying z_0 through the control pathway. This makes it unclear that faithful reconstruction forces the model to internalize semantic and distortion information. Please provide evidence against an identity shortcut: for example, ablate z_0 conditioning (condition only on z_t or on a corrupted/perturbed z_0), measure reconstruction error on distorted inputs, and show that the extracted features change meaningfully when the input degradation changes.
  3. [Section 4.2, Section 4.1] The data provenance for the Controller adaptation is underspecified. Section 4.2 says the Controller is trained 'using targeted VQA datasets,' but does not state whether this adaptation is restricted to the training split of each dataset or performed separately for each evaluation scheme. If the Controller adaptation ever uses the test split or the full dataset before splitting, the intra-dataset and cross-dataset numbers in Tables 1-4 would be leakage-affected. Please specify the exact splits used for Controller adaptation and, ideally, include a nested or per-fold adaptation protocol so the reported averages are unbiased.
  4. [Section 5.2, Figure 7] The claim that the diffusion extractor yields 'clearly separated clusters for different semantic and distortion representations' is supported only by qualitative t-SNE plots on a small synthetic set (10 videos for semantics, 50 for distortion). This is used to justify the semantic/distortion decoupling that underlies the two-branch design. Please provide a quantitative cluster-separation metric (for example, classification accuracy of degradation type or a silhouette score) and a head-to-head comparison with ViT on the same evaluation, so the decoupling claim is not based solely on visual inspection.
minor comments (5)
  1. [Section 4.1] The reference [17] is used for both KoNViD-1k and LIVE-VQC; LIVE-VQC is a different dataset and should be cited with its own reference (e.g., Ghadiyaram et al., 2018).
  2. [References] References [43] and [44] appear to be the same paper, both titled 'High-resolution image synthesis with latent diffusion models,' but they are cited in different contexts and listed with the same venue details; please merge or disambiguate.
  3. [Figure 3(b)] The label 'Imgrestoration vs. Imgreconstruction' contains a typo; it should read 'Image restoration vs. Image reconstruction.'
  4. [Section 3.2, Eq. (10)] In Eq. (10), the subscript 'z0' is typeset with a space as 'z 0'; please fix the formatting.
  5. [Section 5.1] The methods list in Section 5.1 mentions NIQE, TPQI, and SAQI, but these do not appear in any of the result tables; please either include their results or remove them from the list.

Circularity Check

0 steps flagged · score 2.0 of 10

No definitional circularity; reported SRCC/PLCC values are independent held-out measurements, but a minor non-load-bearing self-citation and two validity caveats (target-dataset Controller adaptation and the z0 shortcut in Eq. 10) prevent a clean causal reading.

full rationale

The central claim is empirical rather than definitional. DiffVQA's scores are evaluated on held-out test splits of LIVE-VQC, KoNViD-1k, YouTube-UGC, LSVQ, and KVQ against ground-truth MOS using SRCC/PLCC; no equation constructs the output score from the training target, and the diffusion features are not defined in terms of the VQA labels. The self-citation [4] (UniRestore, overlapping authors) appears in the introduction and related work in support of diffusion-model generalizability and image restoration, but it is accompanied by independent external citations [30,65] and is not the load-bearing evidence for the VQA result, so it does not constitute circularity under the given standards. Two caveats are real but are correctness/validity concerns rather than circular reductions. First, Section 4.2 states that the Controller is trained for 30,000 iterations on 'targeted VQA datasets,' while Table 5's comparison backbones receive only pre-trained initialization; if that adaptation stage sees the same datasets later used for evaluation, the claim that 'all variants are trained under identical conditions' is misleading and the backbone comparison is confounded, but this is unequal training rather than a definitional or statistical reduction. Second, the Controller input in Eq. 10 concatenates the noisy latent z_t with the clean latent z_0; since Eq. 1 defines z_t as a linear function of z_0 and the noise epsilon, the reconstruction loss can be driven near zero by an algebraic shortcut without learning generalizable semantic or distortion features. This undercuts the 'underlying hypothesis' stated in Section 3.2, but the paper explicitly presents that statement as a hypothesis, and the reported benchmark numbers remain independent measurements. No circular step in the paper's own equations or self-citation chain meets the evidentiary bar for a definitional reduction; the modest score reflects the minor self-citation and the confounding adaptation protocol rather than a genuinely circular derivation.

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

The paper's contribution is empirical; it introduces no new physical entities or mathematical axioms. The central claim rests on four assumptions: the transferability of Stable Diffusion features, the reconstruction-to-features hypothesis, the semantic versus distortion branch split, and the temporal modeling advantage of Mamba. The main free parameters are the trained Controller weights and several underreported hyperparameters. The most fragile item is the reconstruction-to-features hypothesis, which is weakened by the clean-latent conditioning.

free parameters (4)
  • Diffusion feature extractor Controller weights = not reported
    The Controller is trained on targeted VQA datasets (Section 4.2). The resulting features are central to the claim, but the weights are not released and the exact training dataset per experiment is not specified.
  • Number of stacked Bidirectional Mamba blocks L in TCAB = not specified
    Figure 6 shows '× L' but L is never defined; this is a required hyperparameter for reproducing the temporal module.
  • Number of DDPM sampling steps at inference = 10
    Chosen for efficiency (Section 4.2); the choice affects the extracted features and is not justified by analysis.
  • Input frame size = 384x384
    Input frames are resized and randomly cropped to 384x384 (Section 4.2); no ablation justifies this resolution.
assumptions (4)
  • domain assumption Stable Diffusion 2.1-base pretrained on large-scale image-text data provides generalizable semantic and distortion features usable for VQA.
    Core premise of the method; Section 3.2 assumes the diffusion prior is valuable for quality assessment.
  • ad hoc to paper Reconstructing identical input frames forces the adapted model to encode the critical semantic and distortion information internally.
    Stated as 'the underlying hypothesis' in Section 3.2; not proven and weakened by conditioning on the clean latent z0.
  • ad hoc to paper The resize branch captures semantics and the random-crop branch captures distortion.
    Design assumption from DisCoVQA [58]; used to route features into separate MLP heads; only qualitative t-SNE support is given.
  • domain assumption Mamba-based state space models capture long-term temporal dependencies better than 3D convolutions for VQA.
    Supported only by a single ablation on LIVE-VQC (Table 6), not by cross-dataset temporal analysis.

how reviews work

0 comments
Cite this review

Pith. "Pith review of DiffVQA: Video Quality Assessment Using Diffusion Feature Extractor." pith.science (2026). https://pith.science/paper/35OQLLSB

@misc{pith2026250503261,
  author       = {Pith},
  title        = {Pith review of: DiffVQA: Video Quality Assessment Using Diffusion Feature Extractor},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/35OQLLSB}},
  note         = {Machine review of arXiv:2505.03261}
}
read the original abstract

Video Quality Assessment (VQA) aims to evaluate video quality based on perceptual distortions and human preferences. Despite the promising performance of existing methods using Convolutional Neural Networks (CNNs) and Vision Transformers (ViTs), they often struggle to align closely with human perceptions, particularly in diverse real-world scenarios. This challenge is exacerbated by the limited scale and diversity of available datasets. To address this limitation, we introduce a novel VQA framework, DiffVQA, which harnesses the robust generalization capabilities of diffusion models pre-trained on extensive datasets. Our framework adapts these models to reconstruct identical input frames through a control module. The adapted diffusion model is then used to extract semantic and distortion features from a resizing branch and a cropping branch, respectively. To enhance the model's ability to handle long-term temporal dynamics, a parallel Mamba module is introduced, which extracts temporal coherence augmented features that are merged with the diffusion features to predict the final score. Experiments across multiple datasets demonstrate DiffVQA's superior performance on intra-dataset evaluations and its exceptional generalization across datasets. These results confirm that leveraging a diffusion model as a feature extractor can offer enhanced VQA performance compared to CNN and ViT backbones.

Figures

Figures reproduced from arXiv: 2505.03261 by the authors.

Figure 1
Figure 1. PLCC and SRCC performance across intra- and cross-dataset evaluations. DiffVQA, represented by red trian￾gles, shows superior performance against established methods like DOVER [61], FAST-VQA [57], TLVQM [22], VIDEVAL [52], and Patch-VQ [66]. We evaluate DiffVQA on the KoNViD-1k [17] dataset using models trained on KoNViD-1k for intra-dataset eval￾uation and LSVQ [66] for cross-dataset evaluation, demonstrating its … view at source ↗
Figure 2
Figure 2. Illustration of Bidirectional Mamba Block. The initial normalization and the final residual are omitted for simplification. z_t by iteratively removing the added noise. This denoising step is achieved by a U-Net denoted by \epsilon _\theta , which is de￾signed to predict the noise component. The sample z_{t-1} at timestep t-1 is derived from z_t as: z_{t-1} = \frac {1}{\sqrt {\alpha _t}} \left (z_t - \frac {1 - \alp… view at source ↗
Figure 3
Figure 3. Adaptation process of the Diffusion Feature Extrac￾tor shown in (a). LAdap is computed between the added noise \epsilon and the predicted noise \epsilon _{\theta } . (b) A similar architecture can be used for image restoration, but we repurpose it for image reconstruction here. (c) During inference, we use \protect \hat {z}_0 , along with the features from Denoising Network at time step t = 0 as the extracted featur… view at source ↗
Figures from the paper (4 more)
Figure 4
Figure 4. Figure 4: Examples of reconstructed results from the adapted diffusion model, along with error maps showing the differ￾ences between the input and reconstructed frames. After adap￾tation, the diffusion model reconstructs the input frames with min￾imal error. each frame to get th…
Figure 5
Figure 5. Figure 5: Architecture of DiffVQA. The diffusion feature extractor extracts semantic and distortion features from video frames, which are enhanced by the DFF, TDM, and FFF modules. The TCAB are also used to capture temporal coherence in parallel. Features extracted from the resi…
Figure 6
Figure 6. Figure 6: Architecture of the Temporal Coherence Augmenta￾tion Block. This block utilizes a 3D convolution block to generate 3D patch embeddings. It employs a bidirectional Mamba block ( [PITH_FULL_IMAGE:figures/full_fig_p006_6.png]
Figure 7
Figure 7. Figure 7: t-SNE visualization of semantic and distortion repre￾sentations by ViT-L/16 and the diffusion feature extractor. datasets, highlighting the effectiveness of using generative priors from a diffusion model for video quality assessment. 5.2. Ablation Study Effectiveness o…

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. VQualA 2025 Challenge on Engagement Prediction for Short Videos: Methods and Results

    cs.CV 2025-09 conditional novelty 4.0 of 10

    A challenge report showing multi-modal models reach SROCC 0.710 in predicting short-video engagement continuation rate, beating a 0.660 baseline.

Reference graph

Works this paper leans on

70 extracted references · 46 canonical work pages · cited by 1 Pith paper

  1. [1]

    Vivit: A video vi- sion transformer

    Anurag Arnab, Mostafa Dehghani, Georg Heigold, Chen Sun, Mario Lu ˇci´c, and Cordelia Schmid. Vivit: A video vi- sion transformer. In ICCV, 2021. 5

    work page 2021

  2. [2]

    Blended diffusion for text-driven editing of natural images

    Omri Avrahami, Dani Lischinski, and Ohad Fried. Blended diffusion for text-driven editing of natural images. In CVPR,

    work page

  3. [3]

    Learning generalized spatial-temporal deep feature representation for no-reference video quality as- sessment

    Baoliang Chen, Lingyu Zhu, Guo Li, Fangbo Lu, Hongfei Fan, and Shiqi Wang. Learning generalized spatial-temporal deep feature representation for no-reference video quality as- sessment. TCSVT, 2021. 6, 7

    work page 2021

  4. [4]

    Unirestore: Unified perceptual and task-oriented image restoration model using diffusion prior

    I Chen, Wei-Ting Chen, Yu-Wei Liu, Yuan-Chun Chiang, Sy-Yen Kuo, Ming-Hsuan Yang, et al. Unirestore: Unified perceptual and task-oriented image restoration model using diffusion prior. arXiv preprint arXiv:2501.13134, 2025. 2

    arXiv 2025

  5. [5]

    Con- trolstyle: Text-driven stylized image generation using diffu- sion priors

    Jingwen Chen, Yingwei Pan, Ting Yao, and Tao Mei. Con- trolstyle: Text-driven stylized image generation using diffu- sion priors. In ACMMM, 2023. 2

    work page 2023

  6. [6]

    Learning phrase representations using rnn encoder-decoder for statistical machine translation

    Kyunghyun Cho. Learning phrase representations using rnn encoder-decoder for statistical machine translation. arXiv preprint arXiv:1406.1078, 2014. 2

    arXiv 2014

  7. [7]

    An image is worth 16x16 words: Transformers for image recognition at scale

    Alexey Dosovitskiy. An image is worth 16x16 words: Transformers for image recognition at scale. arXiv preprint arXiv:2010.11929, 2020. 1, 4, 8

    arXiv 2010

  8. [8]

    Konvid-150k: A dataset for no-reference video qual- ity assessment of videos in-the-wild

    Franz Götz-Hahn, Vlad Hosu, Hanhe Lin, and Dietmar Saupe. Konvid-150k: A dataset for no-reference video qual- ity assessment of videos in-the-wild. IEEE Access, 2021. 1

    work page 2021

Show all 70 references
  1. [9]

    Mamba: Linear-time sequence modeling with selective state spaces

    Albert Gu and Tri Dao. Mamba: Linear-time sequence modeling with selective state spaces. arXiv preprint arXiv:2312.00752, 2023. 3

    2023 arXiv

  2. [10]

    Efficiently modeling long sequences with structured state spaces

    Albert Gu, Karan Goel, and Christopher Ré. Efficiently modeling long sequences with structured state spaces. arXiv preprint arXiv:2111.00396, 2021. 3, 5

    2021 arXiv

  3. [11]

    Combining recurrent, convolutional, and continuous-time models with linear state space layers

    Albert Gu, Isys Johnson, Karan Goel, Khaled Saab, Tri Dao, Atri Rudra, and Christopher Ré. Combining recurrent, convolutional, and continuous-time models with linear state space layers. NeurIPS, 2021. 3

    2021

  4. [12]

    Deep residual learning for image recognition

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

  5. [13]

    Multi-scale representation learning for image restoration with state-space model

    Yuhong He, Long Peng, Qiaosi Yi, Chen Wu, and Lu Wang. Multi-scale representation learning for image restoration with state-space model. arXiv preprint arXiv:2408.10145 ,

    arXiv

  6. [14]

    Pvqm–a perceptual video quality measure

    Andries P Hekstra, John G Beerends, Daniel Ledermann, FE De Caluwe, Stefan Kohler, Robert H Koenen, Samuel Rihs, Matthias Ehrsam, and Denis Schlauss. Pvqm–a perceptual video quality measure. SPIC, 2002. 1

    2002

  7. [15]

    Benchmarking neu- ral network robustness to common corruptions and perturba- tions

    Dan Hendrycks and Thomas Dietterich. Benchmarking neu- ral network robustness to common corruptions and perturba- tions. arXiv preprint arXiv:1903.12261, 2019. 8

    1903 arXiv

  8. [16]

    Cascaded diffusion models for high fidelity image generation

    Jonathan Ho, Chitwan Saharia, William Chan, David J Fleet, Mohammad Norouzi, and Tim Salimans. Cascaded diffusion models for high fidelity image generation. JMLR, 2022. 2

    2022

  9. [17]

    The konstanz natural video database (konvid-1k)

    Vlad Hosu, Franz Hahn, Mohsen Jenadeleh, Hanhe Lin, Hui Men, Tamás Szirányi, Shujun Li, and Dietmar Saupe. The konstanz natural video database (konvid-1k). In QoMEX,

  10. [18]

    Squeeze-and-excitation net- works

    Jie Hu, Li Shen, and Gang Sun. Squeeze-and-excitation net- works. In CVPR, 2018. 5

    2018

  11. [19]

    A new approach to linear filtering and prediction problems

    Rudolph Emil Kalman. A new approach to linear filtering and prediction problems. 1960. 3

    1960

  12. [20]

    Adam: A method for stochastic opti- mization

    Diederik P Kingma. Adam: A method for stochastic opti- mization. arXiv preprint arXiv:1412.6980, 2014. 6

    2014 arXiv

  13. [21]

    Semi-supervised classi- fication with graph convolutional networks

    Thomas N Kipf and Max Welling. Semi-supervised classi- fication with graph convolutional networks. arXiv preprint arXiv:1609.02907, 2016. 3

    2016 arXiv

  14. [22]

    Two-level approach for no-reference con- sumer video quality assessment

    Jari Korhonen. Two-level approach for no-reference con- sumer video quality assessment. TIP, 2019. 1, 2, 6, 7

    2019

  15. [23]

    Blind natural video quality prediction via statistical temporal features and deep spatial features

    Jari Korhonen, Yicheng Su, and Junyong You. Blind natural video quality prediction via statistical temporal features and deep spatial features. In ACMMM, 2020. 6, 7

    2020

  16. [24]

    Blindly assess quality of in-the-wild videos via quality-aware pre-training and motion perception

    Bowen Li, Weixia Zhang, Meng Tian, Guangtao Zhai, and Xianpei Wang. Blindly assess quality of in-the-wild videos via quality-aware pre-training and motion perception. TCSVT, 2022. 6, 7

    2022

  17. [25]

    Quality as- sessment of in-the-wild videos

    Dingquan Li, Tingting Jiang, and Ming Jiang. Quality as- sessment of in-the-wild videos. In ACMMM, 2019. 2, 6, 7

    2019

  18. [26]

    Unified qual- ity assessment of in-the-wild videos with mixed datasets training

    Dingquan Li, Tingting Jiang, and Ming Jiang. Unified qual- ity assessment of in-the-wild videos with mixed datasets training. IJCV, 2021. 2, 7

    2021

  19. [27]

    Which has better visual quality: The clear blue sky or a blurry animal? TMM, 2018

    Dingquan Li, Tingting Jiang, Weisi Lin, and Ming Jiang. Which has better visual quality: The clear blue sky or a blurry animal? TMM, 2018. 7

    2018

  20. [28]

    Videomamba: State space model for efficient video understanding

    Kunchang Li, Xinhao Li, Yi Wang, Yinan He, Yali Wang, Limin Wang, and Yu Qiao. Videomamba: State space model for efficient video understanding. ECCV, 2024. 5

    2024

  21. [29]

    Exploring the ef- fectiveness of video perceptual representation in blind video quality assessment

    Liang Liao, Kangmin Xu, Haoning Wu, Chaofeng Chen, Wenxiu Sun, Qiong Yan, and Weisi Lin. Exploring the ef- fectiveness of video perceptual representation in blind video quality assessment. In ACMMM, 2022. 7

    2022

  22. [30]

    Diff- bir: Toward blind image restoration with generative diffusion prior

    Xinqi Lin, Jingwen He, Ziyan Chen, Zhaoyang Lyu, Bo Dai, Fanghua Yu, Yu Qiao, Wanli Ouyang, and Chao Dong. Diff- bir: Toward blind image restoration with generative diffusion prior. In European Conference on Computer Vision , pages 430–448. Springer, 2024. 2

    2024

  23. [31]

    Scaling and masking: A new paradigm of data sampling for image and video quality assessment

    Yongxu Liu, Yinghui Quan, Guoyao Xiao, Aobo Li, and Jin- jian Wu. Scaling and masking: A new paradigm of data sampling for image and video quality assessment. In AAAI,

  24. [32]

    Vmamba: Visual state space model

    Yue Liu, Yunjie Tian, Yuzhong Zhao, Hongtian Yu, Lingxi Xie, Yaowei Wang, Qixiang Ye, and Yunfan Liu. Vmamba: Visual state space model. arXiv preprint arXiv:2401.10166,

    arXiv

  25. [33]

    Videomam- bapro: A leap forward for mamba in video understanding

    Hui Lu, Albert Ali Salah, and Ronald Poppe. Videomam- bapro: A leap forward for mamba in video understanding. arXiv preprint arXiv:2406.19006, 2024. 3

    2024 arXiv

  26. [34]

    Kvq: Kwai video quality assessment for short-form videos

    Yiting Lu, Xin Li, Yajing Pei, Kun Yuan, Qizhi Xie, Yunpeng Qu, Ming Sun, Chao Zhou, and Zhibo Chen. Kvq: Kwai video quality assessment for short-form videos. In CVPR,

  27. [35]

    Clif-vqa: Enhancing video quality assessment by incorporating high-level semantic information related to human feelings

    Yachun Mi, Yan Shu, Yu Li, Chen Hui, Puchao Zhou, and Shaohui Liu. Clif-vqa: Enhancing video quality assessment by incorporating high-level semantic information related to human feelings. In ACMMM, 2024. 6, 7

    2024

  28. [36]

    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. TIP, 2012. 1

    2012

  29. [37]

    A com- pletely blind video integrity oracle

    Anish Mittal, Michele A Saad, and Alan C Bovik. A com- pletely blind video integrity oracle. TIP, 2015. 2

    2015

  30. [38]

    completely blind

    Anish Mittal, Rajiv Soundararajan, and Alan C Bovik. Mak- ing a “completely blind” image quality analyzer. SPL, 2012. 1, 2, 7

    2012

  31. [39]

    Glide: Towards photorealistic image generation and editing with text-guided diffusion models.arXiv preprint arXiv:2112.10741, 2021

    Alex Nichol, Prafulla Dhariwal, Aditya Ramesh, Pranav Shyam, Pamela Mishkin, Bob McGrew, Ilya Sutskever, and Mark Chen. Glide: Towards photorealistic image generation and editing with text-guided diffusion models.arXiv preprint arXiv:2112.10741, 2021. 2

    2021 arXiv

  32. [40]

    Improved denoising diffusion probabilistic models

    Alexander Quinn Nichol and Prafulla Dhariwal. Improved denoising diffusion probabilistic models. In ICML, 2021. 4, 6

    2021

  33. [41]

    Discrete-time control systems

    Katsuhiko Ogata. Discrete-time control systems. Prentice- Hall, Inc., 1995. 3

    1995

  34. [42]

    Learn- ing transferable visual models from natural language super- vision

    Alec Radford, Jong Wook Kim, Chris Hallacy, Aditya Ramesh, Gabriel Goh, Sandhini Agarwal, Girish Sastry, Amanda Askell, Pamela Mishkin, Jack Clark, et al. Learn- ing transferable visual models from natural language super- vision. In ICML, 2021. 4, 8

    2021

  35. [43]

    High-resolution image syn- thesis with latent diffusion models

    Robin Rombach, Andreas Blattmann, Dominik Lorenz, Patrick Esser, and Björn Ommer. High-resolution image syn- thesis with latent diffusion models. In CVPR, 2022. 2, 3

    2022

  36. [44]

    High-resolution image syn- thesis with latent diffusion models

    Robin Rombach, Andreas Blattmann, Dominik Lorenz, Patrick Esser, and Björn Ommer. High-resolution image syn- thesis with latent diffusion models. In CVPR, 2022. 2, 4, 6

    2022

  37. [45]

    Palette: Image-to-image diffusion models

    Chitwan Saharia, William Chan, Huiwen Chang, Chris Lee, Jonathan Ho, Tim Salimans, David Fleet, and Mohammad Norouzi. Palette: Image-to-image diffusion models. In SIG- GRAPH, 2022. 2

    2022

  38. [46]

    Bidirectional recurrent neural networks

    Mike Schuster and Kuldip K Paliwal. Bidirectional recurrent neural networks. TSP, 1997. 3

    1997

  39. [47]

    Controluda: Controllable diffusion- assisted unsupervised domain adaptation for cross-weather semantic segmentation

    Fengyi Shen, Li Zhou, Kagan Kucukaytekin, Ziyuan Liu, He Wang, and Alois Knoll. Controluda: Controllable diffusion- assisted unsupervised domain adaptation for cross-weather semantic segmentation. arXiv preprint arXiv:2402.06446 ,

    arXiv

  40. [48]

    Simplified state space layers for sequence modeling

    Jimmy TH Smith, Andrew Warrington, and Scott W Linder- man. Simplified state space layers for sequence modeling. arXiv preprint arXiv:2208.04933, 2022. 3, 5

    2022 arXiv

  41. [49]

    Deep unsupervised learning using nonequilibrium thermodynamics

    Jascha Sohl-Dickstein, Eric Weiss, Niru Maheswaranathan, and Surya Ganguli. Deep unsupervised learning using nonequilibrium thermodynamics. In ICML, 2015. 2

    2015

  42. [50]

    Adapool: Expo- nential adaptive pooling for information-retaining downsam- pling

    Alexandros Stergiou and Ronald Poppe. Adapool: Expo- nential adaptive pooling for information-retaining downsam- pling. TIP. 8

  43. [51]

    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 ACMMM, 2022. 6, 7

    2022

  44. [52]

    Ugc-vqa: Benchmarking blind video quality assessment for user generated content

    Zhengzhong Tu, Yilin Wang, Neil Birkbeck, Balu Adsumilli, and Alan C Bovik. Ugc-vqa: Benchmarking blind video quality assessment for user generated content. TIP, 2021. 1, 2, 6, 7

    2021

  45. [53]

    Rapique: Rapid and accurate video quality prediction of user generated content

    Zhengzhong Tu, Xiangxu Yu, Yilin Wang, Neil Birkbeck, Balu Adsumilli, and Alan C Bovik. Rapique: Rapid and accurate video quality prediction of user generated content. OJSP, 2021. 1, 6, 7

    2021

  46. [54]

    Youtube ugc dataset for video compression research

    Yilin Wang, Sasi Inguva, and Balu Adsumilli. Youtube ugc dataset for video compression research. In MMSP, 2019. 5, 7

    2019

  47. [55]

    Rich features for perceptual quality assessment of ugc videos

    Yilin Wang, Junjie Ke, Hossein Talebi, Joong Gon Yim, Neil Birkbeck, Balu Adsumilli, Peyman Milanfar, and Feng Yang. Rich features for perceptual quality assessment of ugc videos. In CVPR, 2021. 1, 6, 7

    2021

  48. [56]

    Cbam: Convolutional block attention module

    Sanghyun Woo, Jongchan Park, Joon-Young Lee, and In So Kweon. Cbam: Convolutional block attention module. In ECCV, 2018. 5

    2018

  49. [57]

    Fast- vqa: Efficient end-to-end video quality assessment with frag- ment sampling

    Haoning Wu, Chaofeng Chen, Jingwen Hou, Liang Liao, Annan Wang, Wenxiu Sun, Qiong Yan, and Weisi Lin. Fast- vqa: Efficient end-to-end video quality assessment with frag- ment sampling. In ECCV, 2022. 1, 3, 6, 7

    2022

  50. [58]

    Discovqa: Temporal distortion-content transformers for video quality assessment

    Haoning Wu, Chaofeng Chen, Liang Liao, Jingwen Hou, Wenxiu Sun, Qiong Yan, and Weisi Lin. Discovqa: Temporal distortion-content transformers for video quality assessment. TCSVT, 2023. 4, 5, 6, 7

    2023

  51. [59]

    Dis- entangling aesthetic and technical effects for video qual- ity assessment of user generated content

    Haoning Wu, Liang Liao, Chaofeng Chen, Jingwen Hou, Annan Wang, Wenxiu Sun, Qiong Yan, and Weisi Lin. Dis- entangling aesthetic and technical effects for video qual- ity assessment of user generated content. arXiv preprint arXiv:2211.04894, 2022. 1, 3, 7

    2022 arXiv

  52. [60]

    Exploring opinion-unaware video quality assessment with semantic affinity criterion

    Haoning Wu, Liang Liao, Jingwen Hou, Chaofeng Chen, Erli Zhang, Annan Wang, Wenxiu Sun, Qiong Yan, and Weisi Lin. Exploring opinion-unaware video quality assessment with semantic affinity criterion. ICME, 2023. 7

    2023

  53. [61]

    Exploring video quality assessment on user generated contents from aesthetic and technical perspectives

    Haoning Wu, Erli Zhang, Liang Liao, Chaofeng Chen, Jing- wen Hou, Annan Wang, Wenxiu Sun, Qiong Yan, and Weisi Lin. Exploring video quality assessment on user generated contents from aesthetic and technical perspectives. In ICCV,

  54. [62]

    Towards explainable in-the-wild video quality assess- ment: a database and a language-prompted approach

    Haoning Wu, Erli Zhang, Liang Liao, Chaofeng Chen, Jing- wen Hou, Annan Wang, Wenxiu Sun, Qiong Yan, and Weisi Lin. Towards explainable in-the-wild video quality assess- ment: a database and a language-prompted approach. In ACMMM, 2023. 6, 7

    2023

  55. [63]

    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. arXiv preprint arXiv:2312.17090, 2023. 7

    2023 arXiv

  56. [64]

    Harnessing the spatial- temporal attention of diffusion models for high-fidelity text- to-image synthesis

    Qiucheng Wu, Yujian Liu, Handong Zhao, Trung Bui, Zhe Lin, Yang Zhang, and Shiyu Chang. Harnessing the spatial- temporal attention of diffusion models for high-fidelity text- to-image synthesis. In ICCV, 2023. 2

    2023

  57. [65]

    Diffir: Efficient diffusion model for image restoration

    Bin Xia, Yulun Zhang, Shiyin Wang, Yitong Wang, Xing- long Wu, Yapeng Tian, Wenming Yang, and Luc Van Gool. Diffir: Efficient diffusion model for image restoration. In ICCV, 2023. 2

    2023

  58. [66]

    Patch-vq:’patching up’the video quality problem

    Zhenqiang Ying, Maniratnam Mandal, Deepti Ghadiyaram, and Alan Bovik. Patch-vq:’patching up’the video quality problem. In CVPR, 2021. 1, 5, 6, 7

    2021

  59. [67]

    Long short-term convolutional transformer for no-reference video quality assessment

    Junyong You. Long short-term convolutional transformer for no-reference video quality assessment. In ACMMM, 2021. 3

    2021

  60. [68]

    Adding conditional control to text-to-image diffusion models

    Lvmin Zhang, Anyi Rao, and Maneesh Agrawala. Adding conditional control to text-to-image diffusion models. In ICCV, 2023. 2, 4

    2023

  61. [69]

    A completely blind video quality evaluator

    Qi Zheng, Zhengzhong Tu, Xiaoyang Zeng, Alan C Bovik, and Yibo Fan. A completely blind video quality evaluator. SPL, 2022. 7

    2022

  62. [70]

    Vision mamba: Efficient visual representation learning with bidirectional state space model

    Lianghui Zhu, Bencheng Liao, Qian Zhang, Xinlong Wang, Wenyu Liu, and Xinggang Wang. Vision mamba: Efficient visual representation learning with bidirectional state space model. arXiv preprint arXiv:2401.09417, 2024. 2, 3, 8

    2024 arXiv

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.