Pith. sign in

REVIEW 3 major objections 4 minor 38 references

TDM: Temporally-Consistent Diffusion Model for All-in-One Real-World Video Restoration

T0 review · 3 major / 4 minor · reviewed 2026-08-10 · deepseek-v4-flash

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

Pith's one-line read A single diffusion model, trained on still images, restores five video degradations.

desk verdict Useful all-in-one video restoration package; the empirical core holds, but the temporal-consistency claims overshoot the evidence and need scoping before publication. read the letter →

arxiv 2501.02269 v1 pith:T7UKM6CL submitted 2025-01-04 cs.CV

Yizhou Li , Zihua Liu , Yusuke Monno , Masatoshi Okutomi This is my paper · ORCID
classification cs.CV
keywords Multi-taskVideoRestorationDiffusionModelsControlNetTemporalConsistencyAll-in-oneDDIMInversionSlidingWindowCross-FrameAttention
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 tries to establish that one diffusion-based model can restore five real-world video degradations—dehazing, deraining, denoising, MP4 compression artifacts, and 4x super-resolution—without any video-level training. It does this by fine-tuning a single ControlNet on single-image pairs while prompting it with the task name, a scheme it calls Task Prompt Guidance. At inference, temporal consistency and content preservation are produced by DDIM inversion plus a new Sliding Window Cross-Frame Attention that averages attention keys and values over neighboring frames. The paper reports the best average perceptual-quality scores (FID/KID) among six compared methods on real-world benchmarks and the best frame-consistency and warping-error scores among diffusion-based baselines. It also states that its temporal consistency still falls short of regression-based video models.

What carries the argument

The load-bearing mechanism is the Sliding Window Cross-Frame Attention (SW-CFA) used with DDIM inversion. In standard self-attention the query, key, and value all come from the current frame's latent; SW-CFA keeps the query local but computes the key and value as averages of the projected latents of the $2N+1$ frames in a window centered on the current frame (Eqs. 7–9), with window radius $N=3$ in all experiments. The averaging is presented as a temporal low-pass filter that suppresses frame-to-frame fluctuations, while DDIM inversion replaces random Gaussian noise with deterministic, spatially coherent noise so that adjacent frames share a stable starting point. Task Prompt Guidance is the other supporting piece: it is the text prompt naming the degradation, and it is what lets a single ControlNet switch between tasks.

What would settle it

Run TDM on a real video with fast camera or object motion and compare warping error and frame consistency for SW-CFA window sizes $N=0$ and $N=3$, while also monitoring FID: the SW-CFA claim predicts lower warping error and higher frame consistency at $N=3$, so a result where consistency improves only at the cost of severe FID degradation, or fails to improve on large-motion clips, would show the temporal-low-pass mechanism is not delivering what the paper claims.

Watch

Extended reading notes

Core claim

The central discovery, on the paper's own terms, is that all-in-one video restoration can be assembled from a pre-trained text-to-image diffusion model rather than trained from scratch or treated as a video problem. A ControlNet that conditions Stable Diffusion on the degraded image is fine-tuned once on mixed single-image restoration datasets; the task name is supplied as the text prompt so the same weights serve all five tasks. Temporal consistency, normally the reason video restoration needs video data, is instead created at inference: DDIM inversion converts each frame's latent into a deterministic noise trajectory that is coherent across frames, and SW-CFA replaces each U-Net self-attention layer with attention whose keys and values are averaged over a sliding window of neighboring frames, which the paper characterizes as temporal low-pass filtering. In experiments, this combination outperforms six regression- and diffusion-based baselines on average perceptual quality and leads the diffusion-based methods on temporal consistency.

Load-bearing premise

The load-bearing premise is that averaging attention keys and values over a sliding window, plus deterministic inversion noise, is enough to keep restored video frames consistent over time without any video-level training or explicit motion handling.

Editorial extensions

If this is right

  • A video restoration system can be extended to a new degradation by fine-tuning on single-image pairs and adding a task-name prompt, with no video data or multi-frame training.
  • Temporal consistency for diffusion-based video restoration is available without optical flow, temporal modules, or video-level training, which removes a major memory and data bottleneck.
  • A single unified model can replace per-task video restoration models for common degradations, reducing deployment and maintenance cost.
  • Because training fits on one GPU using roughly 5,000–6,000 images per task, the approach is accessible to labs without video-scale compute.
  • The paper's reported gap to regression-based temporal consistency defines the next target for diffusion-based video restoration.

Reading between the lines

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

  • An untested implication is that SW-CFA's temporal low-pass behavior has a bandwidth cost: large window sizes may smooth out legitimate fast motion or detail changes, and a sweep over $N$ on high-motion footage would expose that trade-off.
  • Because Task Prompt Guidance leans on the text understanding already inside Stable Diffusion, a natural extension is to prompt with unseen degradation descriptions at inference and test whether the model generalizes to degradations it was never trained on.
  • The paper's own admission that regression-based video models remain more temporally consistent suggests that the next step is coupling SW-CFA with flow-based alignment or a lightweight recurrent module, rather than relying on attention averaging alone.
  • The same single-image training recipe could be applied to other video tasks such as deblurring or low-light enhancement, using the same task prompts and sliding-window attention at inference.
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

3 major / 4 minor

Summary. The paper proposes TDM, a single ControlNet fine-tuned on single-image restoration datasets with task-specific text prompts (Task Prompt Guidance), combined at inference with DDIM inversion and a Sliding Window Cross-Frame Attention (SW-CFA) mechanism. The method is evaluated on five real-world video restoration tasks—dehazing, deraining, denoising, MP4 compression artifact removal, and 4x super-resolution—against regression-based and diffusion-based baselines. The main claims are that one model can handle multiple degradations, that training-free SW-CFA plus DDIM inversion yields temporally consistent video output, and that the method achieves the best average FID/KID among compared methods.

Significance. If the central claims hold, the paper would be a useful practical contribution: a single diffusion-based model for multiple video restoration tasks, trainable on a single GPU, with a training-free temporal-consistency mechanism. The training strategy is simple and reproducible in principle, and the ablation study isolates the contributions of TPG, DDIM inversion, and SW-CFA. The main limitations are the lack of statistical significance analysis, the absence of code and exact prompts, and an internal inconsistency between the abstract's temporal-consistency claim and the conclusion's admission that regression-based methods remain more temporally consistent.

major comments (3)
  1. [Sec. 4 vs. Abstract/Sec. 3.2] The abstract and Section 3.2 state that TDM shows 'superiority' or 'consistently outperforms the others in maintaining temporal consistency,' but Section 4 explicitly concedes that 'our method still falls short of regression-based methods in temporal consistency.' Tables 1 and 2 report WE/FC only for diffusion-based baselines, not for VRT and RVRT, so the temporal-consistency claim is scoped to diffusion methods and is contradicted by the paper's own limitation statement. This needs to be resolved either by adding temporal metrics for regression-based methods in the tables or by clearly re-scoping all claims in the abstract and introduction to 'diffusion-based methods.'
  2. [Sec. 2.3, Eq. (9)] The claim that averaging keys and values over a sliding window 'functions as a temporal low-pass filter' is asserted without derivation. Equation (9) computes softmax(Q_i * mean(K_j)^T) * mean(V_j), which is not equivalent to averaging attention outputs, and for large motions the averaged keys and values can mix spatially incompatible features. The paper should either provide a formal argument or, more practically, include a stress test with large-motion video sequences and an analysis of failure cases; Table 3 alone is not sufficient because it shows SW-CFA slightly degrades FID relative to self-attention on dehazing (73.68 vs. 73.36) and MP4 (100.91 vs. 100.56), indicating that the temporal gain is not cost-free.
  3. [Tables 1-4] All quantitative results are reported from a single run without error bars or repeated-seed variance. Several winning margins are very small (e.g., average KID 4.68 vs. 4.71 for InstructP2P; average FID 84.50 vs. 85.01), and FID/KID on real-world benchmarks can be sensitive to the random seed and sampling steps. The authors should provide standard deviations over multiple runs, or at least multiple seeds, for the main comparisons and ablations, and clarify whether the reported numbers are medians or means.
minor comments (4)
  1. [Sec. 3.1] The exact task prompts used for TPG are never listed in the paper. Since TPG is one of the three core contributions, the prompts should be given explicitly in the text or in an appendix.
  2. [Sec. 3.1] Reference [15] is cited for the denoising dataset, but that reference is a video object segmentation paper; it is unclear how it serves as a denoising dataset. Please clarify the dataset construction and the noise model used for training.
  3. [Table 4] The row labels 'T+I', 'T+S', and 'I+S' are not immediately self-explanatory; define the abbreviations in the table caption or in the text immediately before the table.
  4. [Sec. 4] There is a minor grammar error in the conclusion: 'Experiments across five tasks exhibits proposed TDM’s superior generalization' should be 'exhibit' and 'TDM’s superior generalization.'

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the restoration and temporal-consistency claims are tested against external real-world benchmarks, and no target metric is defined by a fitted constant or self-citation chain.

full rationale

I walked the paper's derivation chain. The training scheme in Sec. 2.2 fine-tunes a ControlNet initialized from the official tile resample model on paired single images with task prompts; no target FID, KID, FC, or WE value is used as a training objective or defined through those constants. The inference mechanism in Sec. 2.3 replaces self-attention with SW-CFA (Eqs. 7-9) and uses DDIM inversion; these are algorithmic choices evaluated empirically in Tables 1-3, not quantities derived from the benchmarks they predict. Hyperparameters such as N=3, 10 inversion steps, and 32 backward steps are hand-set, but hand-setting is not circular unless the selected value is the predicted quantity or is fitted to the evaluation set; the paper gives no evidence of such fitting. The claim that averaging keys and values acts as a temporal low-pass filter is asserted rather than proved, and the Sec. 4 statement that the method 'still falls short of regression-based methods in temporal consistency' conflicts with the abstract's broad wording; these are correctness and overclaim issues, not circularity. No load-bearing self-citations are present: references to Text2Video-zero, ControlVideo, Diff-Plugin, and Stable Diffusion are external prior work used as components or baselines. Therefore no circular step can be exhibited with a specific equation or definitional reduction, and the paper is self-contained in its empirical evaluation against external benchmarks.

Assumptions & free parameters 4 free parameters · 5 assumptions · 0 invented entities

The paper's central contribution is empirical, so no physical or mathematical constants are fitted. The ledger lists hand-set inference and training hyperparameters that the results depend on, plus the domain assumptions behind TPG, SW-CFA, and the chosen metrics. No new physical entities, forces, or conserved quantities are introduced.

free parameters (4)
  • SW-CFA window radius N = N = 3
    Hand-set for all experiments in Sec. 3.1; controls how many neighboring frames are averaged and directly affects temporal consistency and content quality.
  • DDIM inversion and backward step counts = 10 inversion steps, 32 backward steps
    Chosen in Sec. 3.1; affects content preservation, noise determinism, and runtime, with no reported sensitivity sweep.
  • Training schedule and resolution = 25 epochs, batch size 4, learning rate 1e-5, 512x512 patches
    Reported in Sec. 3.1; standard recipe but hand-selected and relevant to reproducing the tables.
  • Task prompt phrasing = "remove the noise", "remove the fog", "recover the high resolution details", "remove the rain"
    TPG in Sec. 2.2 depends on these hand-chosen text prompts; no study shows sensitivity to alternative wording.
assumptions (5)
  • standard math DDIM inversion of the DDIM ODE can be reversed with a small number of steps to give deterministic input noise that preserves content (Eq. 4).
    Inherited from Dhariwal and Song; used in Secs. 2.1 to 2.3 as the basis for content preservation.
  • domain assumption The pre-trained Stable Diffusion text encoder understands task phrases well enough to route the same ControlNet to the correct restoration behavior.
    TPG in Sec. 2.2 relies on SD's zero-shot classification ability; no direct probe of prompt discrimination is reported.
  • domain assumption Replacing self-attention with window-averaged cross-frame attention at inference yields temporal consistency without any video-level training.
    Sec. 2.3 asserts this; no video training or explicit motion model is used.
  • domain assumption FID/KID on reference-free real benchmarks and FC/WE computed with optical flow adequately measure restoration quality and temporal consistency.
    Sec. 3.1 chooses these metrics; no user study, PSNR/SSIM on paired data, or uncertainty estimates are provided.
  • ad hoc to paper Averaging keys and values across a window reduces variance in attention inputs and acts as a temporal low-pass filter.
    Sec. 2.3 states this as a mathematical fact without derivation; it is a paper-specific justification for SW-CFA.

how reviews work

0 comments
Cite this review

Pith. "Pith review of TDM: Temporally-Consistent Diffusion Model for All-in-One Real-World Video Restoration." pith.science (2026). https://pith.science/paper/T7UKM6CL

@misc{pith2026250102269,
  author       = {Pith},
  title        = {Pith review of: TDM: Temporally-Consistent Diffusion Model for All-in-One Real-World Video Restoration},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/T7UKM6CL}},
  note         = {Machine review of arXiv:2501.02269}
}
read the original abstract

In this paper, we propose the first diffusion-based all-in-one video restoration method that utilizes the power of a pre-trained Stable Diffusion and a fine-tuned ControlNet. Our method can restore various types of video degradation with a single unified model, overcoming the limitation of standard methods that require specific models for each restoration task. Our contributions include an efficient training strategy with Task Prompt Guidance (TPG) for diverse restoration tasks, an inference strategy that combines Denoising Diffusion Implicit Models~(DDIM) inversion with a novel Sliding Window Cross-Frame Attention (SW-CFA) mechanism for enhanced content preservation and temporal consistency, and a scalable pipeline that makes our method all-in-one to adapt to different video restoration tasks. Through extensive experiments on five video restoration tasks, we demonstrate the superiority of our method in generalization capability to real-world videos and temporal consistency preservation over existing state-of-the-art methods. Our method advances the video restoration task by providing a unified solution that enhances video quality across multiple applications.

Figures

Figures reproduced from arXiv: 2501.02269 by the authors.

Figure 1
Figure 1. Our Temporally-consistent Diffusion Model (TDM) has two main features: (a) Our model is all-in-one and can restore various real-world video degradation with a single diffusion model under the guidance of task prompts. (b) Our model can generate temporally consistent video frames with better preservation of original contents included in the input video. tensive research has demonstrated their ability to parse and enc… view at source ↗
Figure 2
Figure 2. Overall architecture of our proposed temporally-consistent diffusion model (TDM). 2.2 Training: Task Prompt Guided ControlNet Fine-Tuning with Single-Image Inputs The pre-trained SD model excels at generating high-quality images without degrada￾tion, a feature we strive to maintain, particularly in the face of severe degradations such as sensor noise, low resolution, and haze. To this end, as shown in [PITH_FULL_IM… view at source ↗
Figure 3
Figure 3. Proposed SW-CFA compared with exisiting cross-frame attention. 2.3 Inference: Training-Free Content-Preserved Temporal Consistency for Larger Motion As mentioned earlier, we only use single-image inputs in the training phase, so tem￾poral consistency is addressed through a training-free approach during the inference. Previous studies [14,38] achieve this by replacing self-attention layers in denoising U￾Nets with cr… view at source ↗
Figures from the paper (3 more)
Figure 4
Figure 4. Figure 4: Qualitative comparison with state-of-the-art methods. superior restoration quality with remarkable detail preservation and consistency across different video restoration challenges. Consistency Evaluation. Although diffusion-based methods typically yield images of high…
Figure 5
Figure 5. Figure 5: Consistency comparison (MP4) with other diffusion-based methods [PITH_FULL_IMAGE:figures/full_fig_p011_5.png]
Figure 6
Figure 6. Figure 6: Consistency comparison (SR×4) with other zero-shot cross-frame attention. improvements. It demonstrates a superior ability to maintain uniformity across frames, effectively handling larger motions and providing robust temporal stability [PITH_FULL_IMAGE:figures/full_f…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

38 extracted references · 24 canonical work pages

  1. [1]

    arXiv preprint arXiv:1801.01401 (2018)

    Bi ´nkowski, M., Sutherland, D.J., Arbel, M., Gretton, A.: Demystifying mmd gans. arXiv preprint arXiv:1801.01401 (2018)

    arXiv 2018

  2. [2]

    arXiv preprint arXiv:2311.15127 (2023)

    Blattmann, A., Dockhorn, T., Kulal, S., Mendelevitch, D., Kilian, M., Lorenz, D., Levi, Y ., English, Z., V oleti, V ., Letts, A., et al.: Stable video diffusion: Scaling latent video diffusion models to large datasets. arXiv preprint arXiv:2311.15127 (2023)

    arXiv 2023

  3. [3]

    In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition

    Brooks, T., Holynski, A., Efros, A.A.: Instructpix2pix: Learning to follow image editing instructions. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. pp. 18392–18402 (2023)

    2023

  4. [4]

    Learning Task-Oriented Flows to Mutually Guide Feature Alignment in Synthesized and Real Video Denoising

    Cao, J., Wang, Q., Liang, J., Zhang, Y ., Zhang, K., Timofte, R., Van Gool, L.: Learning task- oriented flows to mutually guide feature alignment in synthesized and real video denoising. arXiv preprint arXiv:2208.11803 (2022) TDM: Temporally-Consistent Diffusion Model 13

    work page Pith review arXiv 2022

  5. [5]

    In: Proceedings of the IEEE/CVF Conference on Com- puter Vision and Pattern Recognition

    Chan, K.C., Zhou, S., Xu, X., Loy, C.C.: Basicvsr++: Improving video super-resolution with enhanced propagation and alignment. In: Proceedings of the IEEE/CVF Conference on Com- puter Vision and Pattern Recognition. pp. 5972–5981 (2022)

    work page 2022

  6. [6]

    arXiv preprint arXiv:2305.15241 (2023)

    Chen, H., Dong, Y ., Wang, Z., Yang, X., Duan, C., Su, H., Zhu, J.: Robust classification via a single diffusion model. arXiv preprint arXiv:2305.15241 (2023)

    arXiv 2023

  7. [7]

    Advances in Neural Information Processing Systems 36 (2024)

    Clark, K., Jaini, P.: Text-to-image diffusion models are zero shot classifiers. Advances in Neural Information Processing Systems 36 (2024)

    work page 2024

  8. [8]

    Advances in Neural Information Processing Systems 34, 8780–8794 (2021)

    Dhariwal, P., Nichol, A.: Diffusion models beat gans on image synthesis. Advances in Neural Information Processing Systems 34, 8780–8794 (2021)

    2021

Show all 38 references
  1. [9]

    IEEE Transactions on Pattern Analysis and Machine Intelligence 43(3), 949–963 (2019)

    Guan, Z., Xing, Q., Xu, M., Yang, R., Liu, T., Wang, Z.: Mfqe 2.0: A new approach for multi- frame quality enhancement on compressed video. IEEE Transactions on Pattern Analysis and Machine Intelligence 43(3), 949–963 (2019)

    2019

  2. [10]

    arXiv preprint arXiv:2307.04725 (2023)

    Guo, Y ., Yang, C., Rao, A., Wang, Y ., Qiao, Y ., Lin, D., Dai, B.: Animatediff: Animate your personalized text-to-image diffusion models without specific tuning. arXiv preprint arXiv:2307.04725 (2023)

    2023 arXiv

  3. [11]

    Advances in Neural Information Processing Systems 30 (2017)

    Heusel, M., Ramsauer, H., Unterthiner, T., Nessler, B., Hochreiter, S.: Gans trained by a two time-scale update rule converge to a local nash equilibrium. Advances in Neural Information Processing Systems 30 (2017)

    2017

  4. [12]

    Advances in Neural Information Processing Systems 33, 6840–6851 (2020)

    Ho, J., Jain, A., Abbeel, P.: Denoising diffusion probabilistic models. Advances in Neural Information Processing Systems 33, 6840–6851 (2020)

    2020

  5. [13]

    In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (2024)

    Ke, B., Obukhov, A., Huang, S., Metzger, N., Daudt, R.C., Schindler, K.: Repurposing diffusion-based image generators for monocular depth estimation. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (2024)

    2024

  6. [14]

    In: Proceedings of the IEEE/CVF International Conference on Computer Vision

    Khachatryan, L., Movsisyan, A., Tadevosyan, V ., Henschel, R., Wang, Z., Navasardyan, S., Shi, H.: Text2video-zero: Text-to-image diffusion models are zero-shot video generators. In: Proceedings of the IEEE/CVF International Conference on Computer Vision. pp. 15954– 15964 (2023)

    2023

  7. [15]

    In: Asian Conference on Computer Vision

    Khoreva, A., Rohrbach, A., Schiele, B.: Video object segmentation with language referring expressions. In: Asian Conference on Computer Vision. pp. 123–141 (2019)

    2019

  8. [16]

    In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition

    Li, B., Liu, X., Hu, P., Wu, Z., Lv, J., Peng, X.: All-in-one image restoration for unknown corruption. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. pp. 17452–17462 (2022)

    2022

  9. [17]

    arXiv preprint arXiv:2201.12288 (2022)

    Liang, J., Cao, J., Fan, Y ., Zhang, K., Ranjan, R., Li, Y ., Timofte, R., Van Gool, L.: Vrt: A video restoration transformer. arXiv preprint arXiv:2201.12288 (2022)

    2022 arXiv

  10. [18]

    Advances in Neural Information Processing Systems 35, 378–393 (2022)

    Liang, J., Fan, Y ., Xiang, X., Ranjan, R., Ilg, E., Green, S., Cao, J., Zhang, K., Timofte, R., Gool, L.V .: Recurrent video restoration transformer with guided deformable attention. Advances in Neural Information Processing Systems 35, 378–393 (2022)

    2022

  11. [19]

    In: Proceedings of the IEEE/CVF International Conference on Computer Vision

    Liu, R., Wu, R., Van Hoorick, B., Tokmakov, P., Zakharov, S., V ondrick, C.: Zero-1-to-3: Zero-shot one image to 3d object. In: Proceedings of the IEEE/CVF International Conference on Computer Vision. pp. 9298–9309 (2023)

    2023

  12. [20]

    arXiv preprint arXiv:2309.03453 (2023)

    Liu, Y ., Lin, C., Zeng, Z., Long, X., Liu, L., Komura, T., Wang, W.: Syncdreamer: Generating multiview-consistent images from a single-view image. arXiv preprint arXiv:2309.03453 (2023)

    2023 arXiv

  13. [21]

    arXiv preprint arXiv:2403.00644 (2024)

    Liu, Y ., Liu, F., Ke, Z., Zhao, N., Lau, R.W.: Diff-plugin: Revitalizing details for diffusion- based low-level tasks. arXiv preprint arXiv:2403.00644 (2024)

    2024 arXiv

  14. [22]

    In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition

    Mokady, R., Hertz, A., Aberman, K., Pritch, Y ., Cohen-Or, D.: Null-text inversion for editing real images using guided diffusion models. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. pp. 6038–6047 (2023)

    2023

  15. [23]

    In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (2019) 14 Y

    Nah, S., Baik, S., Hong, S., Moon, G., Son, S., Timofte, R., Lee, K.M.: Ntire 2019 chal- lenge on video deblurring and super-resolution: Dataset and study. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (2019) 14 Y . Li et al

    2019

  16. [24]

    IEEE Transactions on Pattern Analysis and Machine In- telligence (2023)

    Özdenizci, O., Legenstein, R.: Restoring vision in adverse weather conditions with patch- based denoising diffusion models. IEEE Transactions on Pattern Analysis and Machine In- telligence (2023)

    2023

  17. [25]

    arXiv preprint arXiv:2306.13090 (2023)

    Potlapalli, V ., Zamir, S.W., Khan, S., Khan, F.S.: Promptir: Prompting for all-in-one blind image restoration. arXiv preprint arXiv:2306.13090 (2023)

    2023 arXiv

  18. [26]

    In: Proceedings of the IEEE/CVF International Conference on Computer Vision

    Ren, M., Delbracio, M., Talebi, H., Gerig, G., Milanfar, P.: Multiscale structure guided dif- fusion for image deblurring. In: Proceedings of the IEEE/CVF International Conference on Computer Vision. pp. 10721–10733 (2023)

    2023

  19. [27]

    In: Proceedings of the IEEE/CVF Conference on Com- puter Vision and Pattern Recognition

    Rombach, R., Blattmann, A., Lorenz, D., Esser, P., Ommer, B.: High-resolution image syn- thesis with latent diffusion models. In: Proceedings of the IEEE/CVF Conference on Com- puter Vision and Pattern Recognition. pp. 10684–10695 (2022)

    2022

  20. [28]

    In: Proceedings of ACM SIGGRAPH Conference

    Saharia, C., Chan, W., Chang, H., Lee, C., Ho, J., Salimans, T., Fleet, D., Norouzi, M.: Palette: Image-to-image diffusion models. In: Proceedings of ACM SIGGRAPH Conference. pp. 1–10 (2022)

    2022

  21. [29]

    arXiv preprint arXiv:2010.02502 (2020)

    Song, J., Meng, C., Ermon, S.: Denoising diffusion implicit models. arXiv preprint arXiv:2010.02502 (2020)

    2020 arXiv

  22. [30]

    In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition

    Tassano, M., Delon, J., Veit, T.: Fastdvdnet: Towards real-time deep video denoising with- out flow estimation. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. pp. 1354–1363 (2020)

    2020

  23. [31]

    arXiv preprint arXiv:2308.06571 (2023)

    Wang, J., Yuan, H., Chen, D., Zhang, Y ., Wang, X., Zhang, S.: Modelscope text-to-video technical report. arXiv preprint arXiv:2308.06571 (2023)

    2023 arXiv

  24. [32]

    In: Proceedings of the IEEE/CVF International Conference on Computer Vision

    Wu, J.Z., Ge, Y ., Wang, X., Lei, S.W., Gu, Y ., Shi, Y ., Hsu, W., Shan, Y ., Qie, X., Shou, M.Z.: Tune-a-video: One-shot tuning of image diffusion models for text-to-video generation. In: Proceedings of the IEEE/CVF International Conference on Computer Vision. pp. 7623– 7633 (2023)

    2023

  25. [33]

    In: Proceedings of the IEEE/CVF International Conference on Computer Vision

    Yi, P., Wang, Z., Jiang, K., Jiang, J., Ma, J.: Progressive fusion video super-resolution net- work via exploiting non-local spatio-temporal correlations. In: Proceedings of the IEEE/CVF International Conference on Computer Vision. pp. 3106–3115 (2019)

    2019

  26. [34]

    In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition

    Yue, H., Cao, C., Liao, L., Chu, R., Yang, J.: Supervised raw video denoising with a bench- mark dataset on dynamic scenes. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. pp. 2301–2310 (2020)

    2020

  27. [35]

    In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition

    Yue, Z., Xie, J., Zhao, Q., Meng, D.: Semi-supervised video deraining with dynamical rain generator. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. pp. 642–652 (2021)

    2021

  28. [36]

    In: Proceedings of the IEEE/CVF International Conference on Computer Vision

    Zhang, L., Rao, A., Agrawala, M.: Adding conditional control to text-to-image diffusion models. In: Proceedings of the IEEE/CVF International Conference on Computer Vision. pp. 3836–3847 (2023)

    2023

  29. [37]

    In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition

    Zhang, X., Dong, H., Pan, J., Zhu, C., Tai, Y ., Wang, C., Li, J., Huang, F., Wang, F.: Learning to restore hazy video: A new real-world dataset and a new method. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. pp. 9239–9248 (2021)

    2021

  30. [38]

    arXiv preprint arXiv:2305.13077 (2023)

    Zhang, Y ., Wei, Y ., Jiang, D., Zhang, X., Zuo, W., Tian, Q.: Controlvideo: Training-free controllable text-to-video generation. arXiv preprint arXiv:2305.13077 (2023)

    2023 arXiv

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.