Pith. sign in

REVIEW 4 major objections 5 minor 38 references

VISION-XL: High Definition Video Inverse Problem Solver using Latent Image Diffusion Models

T0 review · 4 major / 5 minor · reviewed 2026-08-12 · deepseek-v4-flash

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

Pith's one-line read Latent diffusion models can solve high-definition video inverse problems without fine-tuning, by pairing a pseudo-batch sampling loop with measurement-derived initialization and a low-pass filter that keeps the latents on the clean…

desk verdict Useful practical advance in video inverse problems with latent diffusion, but the paper's own Table 3 undercuts the broad 'state-of-the-art' claim. read the letter →

arxiv 2412.00156 v4 pith:QYM5FJ4Q submitted 2024-11-29 cs.CV cs.AIcs.LGstat.ML

Taesung Kwon , Jong Chul Ye This is my paper · ORCID
classification cs.CVcs.AIcs.LGstat.ML
keywords videoinverseproblemslatentdiffusionmodelsSDXLpseudo-batchsamplinginversiontemporalconsistencylow-passfilterconjugategradient
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 is trying to establish that latent image diffusion models can solve high-definition video inverse problems—situations where the only measured video is a temporal average of clean frames that are also blurred, downsampled, or partially masked. The proposed framework, VISION-XL, adds two mechanisms to the SDXL latent sampler: pseudo-batch sampling, which denoises one frame at a time to fit on a single GPU, and pseudo-batch inversion, which initializes the latent path from the measurement rather than from noise. It also applies a scheduled low-pass filter before re-encoding the data-consistent batch into latent space, preventing the VAE's encoder-decoder round trip from kicking the reconstruction off the clean image manifold. The paper reports state-of-the-art perceptual quality (FVD) across multiple aspect ratios and several inverse problems, at resolutions above 1280×720 in under 6 seconds per frame on a single NVIDIA 4090 GPU. If true, this means progress in latent image generation transfers directly to video restoration without task-specific training.

What carries the argument

The load-bearing mechanism is the alternating loop of latent denoising and pixel-space data consistency, glued by a scheduled low-pass filter. Each iteration: Tweedie-denoise each latent frame separately (pseudo-batch), decode, run an $l$-step conjugate-gradient refinement against the measurement $Y$, apply a low-pass filter $h_{\sigma_t}$ with width $\sigma_t = \lambda\sqrt{1-\bar\alpha_t}$, re-encode, and re-noise with batch-consistent noise. The filter width scales with the noise level and vanishes as $t\to 0$, so early iterations suppress the high-frequency artifacts that VAE re-encoding introduces, while final iterations preserve detail. Pseudo-batch inversion sets the starting latent $z_\tau$ by encoding the measurement and running DDIM inversion from $z_0$.

What would settle it

Run the method with and without the low-pass filter on a degradation whose spectrum is not low-frequency-dominated, such as high-frequency-patterned blur or a non-smooth mask, and compare FVD and PSNR. If removing the filter does not degrade the output on such a task, or if the optimal filter width must be re-tuned per task, the central mechanism is not general as claimed.

Watch

Extended reading notes

Core claim

The central claim is that the failures of latent diffusion models in video inverse problems are not inherent to the latent space but are fixable by two changes to the sampling loop. First, instead of generating a batch of frames from synchronized noise, the latents are initialized by encoding the measured frames and DDIM-inverting them to a mid-noise timestep. Second, after enforcing data consistency on the decoded batch with conjugate-gradient updates, the batch is low-pass filtered with a cutoff that shrinks with the noise scale, then re-encoded. Together these keep the latents on the clean manifold and yield temporally consistent reconstructions, turning SDXL into a state-of-the-art video inverse solver. The method is train-free, works across degradation types, and supports landscape, vertical, and square aspect ratios.

Load-bearing premise

The whole performance gain depends on the heuristic scheduled low-pass filter: if the chosen filter width $\sigma_t = 2\sqrt{1-\bar\alpha_t}$ fails to generalize to a new degradation, aspect ratio, or model version without re-tuning, the claimed advantage over pixel-space methods disappears.

Editorial extensions

If this is right

  • Any latent image diffusion model can be turned into a video inverse solver without fine-tuning, so advances in text-to-image models transfer directly to video restoration.
  • Computational cost scales with a single frame rather than a large batch, making HD video reconstruction feasible on a single consumer GPU.
  • Because the solver avoids optical flow, it does not suffer from flow-estimation failures under severe degradation and can handle spatio-temporal degradations that optical-flow-based methods cannot address.
  • Frame averaging can be combined with deblurring, super-resolution, or inpainting in one forward model and solved in the same alternating loop.
  • The approach inherently supports multiple aspect ratios, including landscape, vertical, and square formats, without reconfiguration.
  • The reported efficiency (under 6 s/frame at 1280×768 on a 4090) suggests the method is practical for interactive or near-real-time use in video editing pipelines.

Reading between the lines

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

  • The success of the low-pass filter suggests a general recipe for latent inverse solvers: any operation that pushes the latent representation off the training manifold—such as aggressive data-consistency updates, editing, or classifier guidance—might benefit from a scale-matched frequency truncation before re-encoding, extending beyond video to high-resolution single-image restoration.
  • The method relies on a single tuned width parameter $\lambda=2$; whether that cutoff can be derived from the noise schedule or adapted per degradation is an open question, and if it requires per-task tuning the robustness across unseen degradations is uncertain.
  • Since the initialization is only DDIM inversion of the measurement frames, a natural test is whether a stronger inversion (for example, with guidance or multiple inversion steps) further improves temporal consistency, with the supplementary material already sketching a two-round blind-deblurring extension.
  • The reported speed implies that diffusion-based restoration could move from offline batch processing into interactive tools, but this depends on the low-pass filter and inversion remaining stable at higher frame counts than the tested 25 frames.
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 presents VISION-XL, a plug-and-play solver for video inverse problems built on SDXL. The algorithm initializes latents by DDIM-inverting the encoded measurement, denoises frame-wise via Tweedie, enforces data consistency with l-step CG in pixel space, applies a scheduled low-pass filter before re-encoding, and re-noises to continue the diffusion loop. Experiments cover six degradations (Deblur, SR, Inpaint, and their '+' frame-averaging variants) on DAVIS and Pexels at three aspect ratios, with ablations for initialization timestep, CG steps, and low-pass filter strength. The paper claims state-of-the-art reconstruction, multi-aspect support, and under 6 seconds per frame on one NVIDIA 4090 GPU.

Significance. If the results held, VISION-XL would be a meaningful step: it is the first latent-diffusion video inverse solver in the comparison set, it handles memory-intensive SDXL via pseudo-batch sampling, and pseudo-batch inversion provides large FVD gains over random or batch-synchronized noise initialization. The ablations are informative and the efficiency numbers in Table 4 are concrete and useful. However, the central state-of-the-art claim is not supported by the paper's own Table 3, because SVI beats VISION-XL on every metric for DAVIS Inpaint+ and DAVIS Inpaint. The contribution remains worthwhile, but the claims need substantial narrowing before publication.

major comments (4)
  1. [Section 4.2, Table 3] The abstract and Section 4.2 claim 'state-of-the-art' reconstruction and state that the method 'consistently outperforms baseline approaches' across spatio-temporal inverse problems, but Table 3 directly contradicts this. On DAVIS Inpaint+, SVI achieves FVD 208.6 vs Ours 241.1, LPIPS 0.238 vs 0.242, PSNR 29.60 vs 28.81, and SSIM 0.848 vs 0.815; on DAVIS Inpaint, SVI wins all four metrics as well. The claim must be narrowed to the tasks and datasets where the method actually wins (e.g., Deblur+ and SR+), and the sentence about a 'significant reduction in FVD compared to the runner-up across all datasets' should be removed or qualified.
  2. [Section 3, Step 4, Eq. (7)] The scheduled low-pass filter is a load-bearing component: Table 7 shows that removing it increases FVD by roughly 30 points and lowers PSNR by about 1 dB. Yet h_sigma_t is never defined (kernel shape, support, or boundary treatment are all missing), and no derivation is given for why sigma_t = lambda * sqrt(1 - alpha_bar_t) keeps re-encoded latents on the clean manifold. The heuristic is validated only on SR+ Pexels landscape and then transferred to every other task and aspect ratio. Please specify the filter explicitly and provide either a derivation or a multi-task validation; as written, the method is not fully reproducible at this step.
  3. [Section 4.3, Tables 5-7] The hyperparameters tau = 0.3T, lambda = 2, and l = 10 are selected by ablations on a single task (SR+, Pexels landscape) and then applied to every other degradation, aspect ratio, and dataset, with no error bars, confidence intervals, or repeated-run statistics. Given the sensitivity shown in Table 6 (l = 5 vs l = 10 changes FVD by about 56 points) and Table 7 (lambda = 1 vs lambda = 2 changes LPIPS and PSNR), the broad-applicability claim needs either additional evidence across tasks or a more conservative statement of the method's tuning requirements.
  4. [Section 4.1, Baselines] The state-of-the-art claim is also weakened by the comparison set: Warped Diffusion is not evaluated, and DiffIR2VR is run only on SR, so the comparison against prior video inverse problem solvers rests largely on SVI, which wins on the inpainting tasks in Table 3. Please either add the missing baseline or explicitly state that the comparison covers only the methods that could be run; the current wording implies a broader comparison than the experiments support.
minor comments (5)
  1. [Algorithm 1] The loop header 'for t = tau : 2 do' is ambiguous; please write it as a descending loop, e.g., 'for t = tau downto 2'.
  2. [Section 3, Step 4] The convolution in Eq. (7) does not specify boundary conditions or kernel support; adding these details would improve reproducibility.
  3. [Section 7, Algorithm 2] The blind video deblurring extension is presented only qualitatively; a quantitative evaluation of the two-round PSF refinement would make the extension more convincing.
  4. [General] No code release is mentioned; given the number of unspecified details, releasing the inference code would substantially help reproducibility.
  5. [Section 4.2] The statement that DiffIR2VR artifacts are 'likely due to inaccuracies in optical flow estimation' is speculative; either provide supporting evidence or soften the wording.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: VISION-XL is an empirically benchmarked algorithm; no derivation step reduces to its inputs by construction.

full rationale

The paper makes no claim that its outputs are derived from first principles; it proposes an algorithm and evaluates it against external baselines. Pseudo-batch inversion initializes latents from the measurement via z0 = Eθ(Y), zτ = DDIM^{-1}(z0), which is an initialization, not a fitted parameter that encodes the target reconstruction. The iterative updates (Tweedie denoising, CG data consistency, low-pass filtered encoding, renoising) are explicit and are not defined in terms of the reported metrics. Hyperparameters τ, λ, and l are selected by ablation on one task and then held fixed across evaluations; this is standard empirical practice and does not make the reported comparisons equivalent to the fitted values by construction. Self-citations to SVI [13] supply the baseline framework and batch-consistent noise idea, but the central contribution—latent-space pseudo-batch sampling and inversion with SDXL—is independently compared against SVI, DiffIR2VR, and ADMM-TV. No uniqueness theorem, definitional identity, or renamed known result forces the outputs to equal the inputs. Claims about state-of-the-art performance are contestable from Table 3, but that is a correctness/scope issue, not circularity.

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

The method builds on a pretrained SDXL prior and introduces no new physical or conceptual entities. It does rely on four hand-chosen hyperparameters and one ad hoc low-pass filter assumption that is central to controlling VAE drift. These are inputs the reader must accept without independent derivation.

free parameters (4)
  • initialization timestep tau = 0.3T
    Chosen by ablation on Pexels landscape SR+ (Table 5). It controls how much measurement information is carried into sampling and is tuned to a target dataset.
  • low-pass filter strength lambda = 2
    Chosen by ablation (Table 7). The schedule sigma_t = lambda * sqrt(1 - alpha_bar_t) is a heuristic without derivation.
  • CG update steps l = 10
    Chosen by ablation (Table 6). Too few steps hurt data consistency; too many degrade perceptual quality.
  • total diffusion steps T = 25
    Set for all experiments with no derivation or ablation table. It is a hand-chosen hyperparameter for SDXL sampling.
assumptions (4)
  • standard math DDIM inversion and Tweedie's formula provide valid approximate posterior latents at each step in latent space.
    Used in Steps 1-2, Eqs. (2)-(4). These are standard tools from Song et al. and Efron.
  • domain assumption SDXL pretrained on images is a sufficient video-frame prior under null-text conditioning.
    Central to the method. SDXL is trained on images, not videos, and the paper validates this only empirically on selected datasets.
  • domain assumption Pixel-space CG data consistency in Eq. (6) is accurate enough without gradient backpropagation through the diffusion model.
    The l-step Krylov subspace approximation is taken from DDS; its sufficiency for latent diffusion with VAE encoding is assumed rather than proven.
  • ad hoc to paper The low-pass filter h_sigma_t with sigma_t = lambda * sqrt(1 - alpha_bar_t) prevents VAE drift and keeps re-encoded latents near the clean manifold.
    No derivation is given; only ablation Table 7 supports it. The filter kernel is not specified, making the assumption hard to reproduce.

how reviews work

0 comments
Cite this review

Pith. "Pith review of VISION-XL: High Definition Video Inverse Problem Solver using Latent Image Diffusion Models." pith.science (2026). https://pith.science/paper/QYM5FJ4Q

@misc{pith2026241200156,
  author       = {Pith},
  title        = {Pith review of: VISION-XL: High Definition Video Inverse Problem Solver using Latent Image Diffusion Models},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/QYM5FJ4Q}},
  note         = {Machine review of arXiv:2412.00156}
}
read the original abstract

In this paper, we propose a novel framework for solving high-definition video inverse problems using latent image diffusion models. Building on recent advancements in spatio-temporal optimization for video inverse problems using image diffusion models, our approach leverages latent-space diffusion models to achieve enhanced video quality and resolution. To address the high computational demands of processing high-resolution frames, we introduce a pseudo-batch consistent sampling strategy, allowing efficient operation on a single GPU. Additionally, to improve temporal consistency, we present pseudo-batch inversion, an initialization technique that incorporates informative latents from the measurement. By integrating with SDXL, our framework achieves state-of-the-art video reconstruction across a wide range of spatio-temporal inverse problems, including complex combinations of frame averaging and various spatial degradations, such as deblurring, super-resolution, and inpainting. Unlike previous methods, our approach supports multiple aspect ratios (landscape, vertical, and square) and delivers HD-resolution reconstructions (exceeding 1280x720) in under 6 seconds per frame on a single NVIDIA 4090 GPU.

Figures

Figures reproduced from arXiv: 2412.00156 by the authors.

Figure 1
Figure 1. Representative video reconstruction by VISION-XL: SR+ (frame averaging with [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Illustration of VISION-XL sampling at timestep [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. Geometric illustration of the sampling path evolution. [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figures from the paper (4 more)
Figure 4
Figure 4. Figure 4: Qualitative evaluation of solving spatio-temporal inverse problems across DAVIS, Pexels dataset with multiple aspect ratios. [PITH_FULL_IMAGE:figures/full_fig_p006_4.png]
Figure 5
Figure 5. Figure 5: Qualitative evaluation of SR (×4) performance across multiple aspect ratios (landscape, vertical). DiffIR2VR often produces unwanted artifacts in the background (red and blue boxes), while SVI inaccurately restores intensity (red box), leading to frame-wise fluctuation…
Figure 6
Figure 6. Figure 6: Ablation study on the effect of the initializations. [PITH_FULL_IMAGE:figures/full_fig_p008_6.png]
Figure 8
Figure 8. Figure 8: Qualitative comparison of video deblurring results on the GoPro test dataset [ [PITH_FULL_IMAGE:figures/full_fig_p012_8.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

38 extracted references · 23 canonical work pages

  1. [1]

    Improving diffusion models for inverse prob- lems using manifold constraints

    Hyungjin Chung, Byeongsu Sim, Dohoon Ryu, and Jong Chul Ye. Improving diffusion models for inverse prob- lems using manifold constraints. Advances in Neural Infor- mation Processing Systems, 35:25683–25696, 2022. 2, 4

    work page 2022

  2. [2]

    Diffusion pos- terior sampling for general noisy inverse problems

    Hyungjin Chung, Jeongsol Kim, Michael Thompson Mc- cann, Marc Louis Klasky, and Jong Chul Ye. Diffusion pos- terior sampling for general noisy inverse problems. In Inter- national Conference on Learning Representations, 2023. 2, 4

    work page 2023

  3. [3]

    Prompt-tuning latent diffusion mod- els for inverse problems

    Hyungjin Chung, Jong Chul Ye, Peyman Milanfar, and Mauricio Delbracio. Prompt-tuning latent diffusion mod- els for inverse problems. arXiv preprint arXiv:2310.01110,

    arXiv

  4. [4]

    De- composed diffusion sampler for accelerating large-scale in- verse problems

    Hyungjin Chung, Suhyeon Lee, and Jong Chul Ye. De- composed diffusion sampler for accelerating large-scale in- verse problems. In The Twelfth International Conference on Learning Representations, 2024. 2, 3, 4

    work page 2024

  5. [5]

    Warped Diffusion: Solving Video Inverse Problems with Image Diffusion Models

    Giannis Daras, Weili Nie, Karsten Kreis, Alex Dimakis, Morteza Mardani, Nikola Borislavov Kovachki, and Arash Vahdat. Warped diffusion: Solving video inverse prob- lems with image diffusion models. arXiv preprint arXiv:2410.16152, 2024. 2, 3, 5

    work page Pith review arXiv

  6. [6]

    Diffusion models beat gans on image synthesis

    Prafulla Dhariwal and Alexander Nichol. Diffusion models beat gans on image synthesis. Advances in neural informa- tion processing systems, 34:8780–8794, 2021. 1, 2, 3

    work page 2021

  7. [7]

    Tweedie’s formula and selection bias

    Bradley Efron. Tweedie’s formula and selection bias. Jour- nal of the American Statistical Association, 106(496):1602– 1614, 2011. 3, 4

    work page 2011

  8. [8]

    An image is worth one word: Personalizing text-to- image generation using textual inversion

    Rinon Gal, Yuval Alaluf, Yuval Atzmon, Or Patash- nik, Amit H Bermano, Gal Chechik, and Daniel Cohen- Or. An image is worth one word: Personalizing text-to- image generation using textual inversion. arXiv preprint arXiv:2208.01618, 2022. 1

    arXiv 2022

Show all 38 references
  1. [9]

    Denoising diffu- sion probabilistic models

    Jonathan Ho, Ajay Jain, and Pieter Abbeel. Denoising diffu- sion probabilistic models. Advances in Neural Information Processing Systems, 33:6840–6851, 2020. 1, 2, 4

    2020

  2. [10]

    Elucidating the design space of diffusion-based generative models

    Tero Karras, Miika Aittala, Timo Aila, and Samuli Laine. Elucidating the design space of diffusion-based generative models. Advances in neural information processing systems, 35:26565–26577, 2022. 1, 4

    2022

  3. [11]

    Denoising diffusion restoration models

    Bahjat Kawar, Michael Elad, Stefano Ermon, and Jiaming Song. Denoising diffusion restoration models. Advances in Neural Information Processing Systems, 35:23593–23606,

  4. [12]

    Regularization by texts for latent diffusion inverse solvers

    Jeongsol Kim, Geon Yeong Park, Hyungjin Chung, and Jong Chul Ye. Regularization by texts for latent diffusion inverse solvers. arXiv preprint arXiv:2311.15658, 2023. 3

    2023 arXiv

  5. [13]

    Solving video inverse problems using image diffusion models

    Taesung Kwon and Jong Chul Ye. Solving video inverse problems using image diffusion models. In The Thir- teenth International Conference on Learning Representa- tions, 2025. 2, 3, 4, 5, 6, 7, 1

    2025

  6. [14]

    Visual decoding and reconstruction via eeg embeddings with guided diffusion

    Dongyang Li, Chen Wei, Shiying Li, Jiachen Zou, and Quanying Liu. Visual decoding and reconstruction via eeg embeddings with guided diffusion. arXiv preprint arXiv:2403.07721, 2024. 1

    2024 arXiv

  7. [15]

    Deep multi-scale convolutional neural network for dynamic scene deblurring

    Seungjun Nah, Tae Hyun Kim, and Kyoung Mu Lee. Deep multi-scale convolutional neural network for dynamic scene deblurring. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 3883–3891,

  8. [16]

    Improved denoising diffusion probabilistic models

    Alexander Quinn Nichol and Prafulla Dhariwal. Improved denoising diffusion probabilistic models. In International Conference on Machine Learning, pages 8162–8171. PMLR,

  9. [17]

    Sdxl: Improving latent diffusion mod- els for high-resolution image synthesis

    Dustin Podell, Zion English, Kyle Lacey, Andreas Blattmann, Tim Dockhorn, Jonas M ¨uller, Joe Penna, and Robin Rombach. Sdxl: Improving latent diffusion mod- els for high-resolution image synthesis. arXiv preprint arXiv:2307.01952, 2023. 1, 2, 3, 5

    2023 arXiv

  10. [18]

    The 2017 davis challenge on video object segmentation.arXiv preprint arXiv:1704.00675, 2017

    Jordi Pont-Tuset, Federico Perazzi, Sergi Caelles, Pablo Ar- bel´aez, Alex Sorkine-Hornung, and Luc Van Gool. The 2017 davis challenge on video object segmentation.arXiv preprint arXiv:1704.00675, 2017. 5

    2017 arXiv

  11. [19]

    High-resolution image synthesis with latent diffusion models

    Robin Rombach, Andreas Blattmann, Dominik Lorenz, Patrick Esser, and Bj ¨orn Ommer. High-resolution image synthesis with latent diffusion models. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 10684–10695, 2022. 1, 2, 3

    2022

  12. [20]

    Solving linear inverse problems provably via posterior sampling with latent diffusion models

    Litu Rout, Negin Raoof, Giannis Daras, Constantine Cara- manis, Alex Dimakis, and Sanjay Shakkottai. Solving linear inverse problems provably via posterior sampling with latent diffusion models. Advances in Neural Information Process- ing Systems, 36, 2024. 3

    2024

  13. [21]

    Denois- ing diffusion implicit models

    Jiaming Song, Chenlin Meng, and Stefano Ermon. Denois- ing diffusion implicit models. In International Conference on Learning Representations, 2021. 1, 2, 4

    2021

  14. [22]

    Pseudoinverse-guided diffusion models for inverse problems

    Jiaming Song, Arash Vahdat, Morteza Mardani, and Jan Kautz. Pseudoinverse-guided diffusion models for inverse problems. In International Conference on Learning Repre- sentations, 2023. 2, 3

    2023

  15. [23]

    Score-based generative modeling through stochastic differential equa- tions

    Yang Song, Jascha Sohl-Dickstein, Diederik P Kingma, Ab- hishek Kumar, Stefano Ermon, and Ben Poole. Score-based generative modeling through stochastic differential equa- tions. In International Conference on Learning Represen- tations, 2020. 1, 2

    2020

  16. [24]

    Evaluating a synthetic image dataset gener- ated with stable diffusion

    Andreas St ¨ockl. Evaluating a synthetic image dataset gener- ated with stable diffusion. In International Congress on In- formation and Communication Technology, pages 805–818. Springer, 2023. 1

    2023

  17. [25]

    High-resolution image re- construction with latent diffusion models from human brain activity

    Yu Takagi and Shinji Nishimoto. High-resolution image re- construction with latent diffusion models from human brain activity. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages 14453– 14463, 2023. 1

    2023

  18. [26]

    Raft: Recurrent all-pairs field transforms for optical flow

    Zachary Teed and Jia Deng. Raft: Recurrent all-pairs field transforms for optical flow. In Computer Vision–ECCV 2020: 16th European Conference, Glasgow, UK, August 23– 28, 2020, Proceedings, Part II 16, pages 402–419. Springer,

    2020

  19. [27]

    Fvd: A new metric for video generation

    Thomas Unterthiner, Sjoerd van Steenkiste, Karol Kurach, Rapha¨el Marinier, Marcin Michalski, and Sylvain Gelly. Fvd: A new metric for video generation. 2019. 2, 5 9

    2019

  20. [28]

    Zero-shot image restoration using denoising diffusion null-space model

    Yinhuai Wang, Jiwen Yu, and Jian Zhang. Zero-shot image restoration using denoising diffusion null-space model. In The Eleventh International Conference on Learning Repre- sentations, 2023. 2, 3

    2023

  21. [29]

    Image quality assessment: from error visibility to structural similarity

    Zhou Wang, Alan C Bovik, Hamid R Sheikh, and Eero P Si- moncelli. Image quality assessment: from error visibility to structural similarity. IEEE transactions on image processing, 13(4):600–612, 2004. 5

    2004

  22. [30]

    Dreamclean: Restoring clean image using deep diffusion prior

    Jie Xiao, Ruili Feng, Han Zhang, Zhiheng Liu, Zhantao Yang, Yurui Zhu, Xueyang Fu, Kai Zhu, Yu Liu, and Zheng- Jun Zha. Dreamclean: Restoring clean image using deep diffusion prior. In The Twelfth International Conference on Learning Representations, 2024. 2, 3

    2024

  23. [31]

    Gmflow: Learning optical flow via global matching

    Haofei Xu, Jing Zhang, Jianfei Cai, Hamid Rezatofighi, and Dacheng Tao. Gmflow: Learning optical flow via global matching. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 8121–8130,

  24. [32]

    Diffusion probabilistic model made slim

    Xingyi Yang, Daquan Zhou, Jiashi Feng, and Xinchao Wang. Diffusion probabilistic model made slim. In Proceedings of the IEEE/CVF Conference on computer vision and pattern recognition, pages 22552–22562, 2023. 3, 4

    2023

  25. [33]

    Diffir2vr- zero: Zero-shot video restoration with diffusion-based image restoration models

    Chang-Han Yeh, Chin-Yang Lin, Zhixiang Wang, Chi-Wei Hsiao, Ting-Hsuan Chen, and Yu-Lun Liu. Diffir2vr- zero: Zero-shot video restoration with diffusion-based image restoration models. arXiv preprint arXiv:2407.01519, 2024. 2, 3, 5, 6, 7, 1

    2024

  26. [34]

    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 Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 3836–3847, 2023. 1

    2023

  27. [35]

    The unreasonable effectiveness of deep features as a perceptual metric

    Richard Zhang, Phillip Isola, Alexei A Efros, Eli Shechtman, and Oliver Wang. The unreasonable effectiveness of deep features as a perceptual metric. In CVPR, 2018. 5 10 VISION-XL: High Definition Video Inverse Problem Solver using Latent Image Diffusion Models Supplementary Material

    2018

  28. [36]

    Implementation of Comparative Methods SVI [13]

    Experimental details 6.1. Implementation of Comparative Methods SVI [13]. For SVI, we use the official implementation 3. Specifically, we utilize the same pre-trained image diffu- sion model, the unconditional ADM [6]. Following the pro- tocol described in [13], we set the par...

  29. [37]

    Here, we provide an example application of our method to blind video de- blurring, showing its potential as a general framework for solving blind video inverse problems

    Extension to blind video inverse problems Our method can be extended to address blind video inverse problems, such as blind video deblurring, demonstrated us- ing the widely-used GoPro dataset [15]. Here, we provide an example application of our method to blind video de- blurr...

  30. [38]

    The page features 36 paired visualizations of mea- surements and reconstructions across various aspect ratios and degradation types

    Comprehensive visualizations For an in-depth understanding of the experimental results, we provide video visualizations on our anonymous project page5. The page features 36 paired visualizations of mea- surements and reconstructions across various aspect ratios and degradation...

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.