pith. sign in

arxiv: 2606.01315 · v1 · pith:LJ6X3D52new · submitted 2026-05-31 · 💻 cs.CV

DeblurNVS: Geometric Latent Diffusion for Novel View Synthesis from Sparse Motion-Blurred Images

Changyue Shi , Wangbo Yu , Chaoran Feng , Li Yuan This is my paper

Pith reviewed 2026-06-28 17:26 UTC · model grok-4.3

classification 💻 cs.CV
keywords novel view synthesismotion blurlatent diffusiongeometric representationssparse viewsdeblurringneural rendering3D reconstruction
0
0 comments X

The pith

DeblurNVS restores geometric representations from sparse motion-blurred images via latent diffusion to synthesize sharp novel views without per-scene optimization.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper establishes a framework for novel view synthesis that starts from motion-blurred and sparse input images rather than clean ones. It first uses latent diffusion to recover intermediate geometric representations that preserve structure and cross-view correspondences, then feeds those plus target camera poses into a synthesis stage that outputs a sharp RGB image. This approach is trained once on a large dataset of synthesized blur and runs directly at inference time. A reader would care because real captures frequently contain blur from camera motion or long exposures, breaking the clean-image assumption of most current NVS pipelines and forcing slow per-scene fitting in prior blur-aware work.

Core claim

DeblurNVS restores the intermediate geometric representations needed for multi-view reasoning from blurred inputs, enabling the synthesis of high-fidelity novel views directly from sparse motion-blurred images without requiring per-scene optimization. The restored representations are combined with target camera information to synthesize the target-view representation and reconstruct a sharp RGB novel view.

What carries the argument

Geometric latent diffusion that restores intermediate geometric representations from motion-blurred observations to recover structure and correspondence cues.

If this is right

  • The method produces perceptually sharper and structurally more stable novel views than existing baselines on synthetic motion-blur benchmarks.
  • It generalizes from the synthetic training distribution to real motion-blurred scenes.
  • It enables efficient sparse-view synthesis by removing the need for costly per-scene optimization at test time.
  • A large-scale motion-blurred NVS dataset constructed via interpolation-based finite-exposure blur synthesis supports the training.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • If geometric restoration proves robust, the same latent-diffusion step could be adapted to other common degradations such as noise or low-light conditions that also weaken multi-view correspondences.
  • The separation of restoration and synthesis stages suggests that future work could swap in different diffusion backbones or geometry encoders without retraining the entire pipeline.
  • Success on real scenes would imply that large synthetic blur datasets can substitute for hard-to-collect paired real blurred-and-sharp multi-view data in practical capture settings.

Load-bearing premise

Restoring intermediate geometric representations from blurred inputs is sufficient to recover reliable structure and correspondence cues needed for accurate multi-view synthesis.

What would settle it

A side-by-side evaluation on real motion-blurred scenes where novel views from DeblurNVS are compared against both clean-image baselines and ground-truth sharp views of the same scenes, checking whether structural stability and sharpness improve measurably.

Figures

Figures reproduced from arXiv: 2606.01315 by Changyue Shi, Chaoran Feng, Li Yuan, Wangbo Yu.

Figure 1
Figure 1. Figure 1: We propose DeblurNVS, a novel framework that synthesizes sharp [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Visualization of DL3DV-10K-Blur. We show examples of preprocessed sharp–blur image pairs from our dataset. pose-conditioned diffusion models for view generation, as in GeNVS [50], Zero-1-to-3 [51], ZeroNVS [52], and Reconfu￾sion [53]. More recent works introduce video diffusion priors to improve view consistency and camera controllability. In par￾ticular, ViewCrafter [3] combines coarse point-based geometr… view at source ↗
Figure 3
Figure 3. Figure 3: Pipeline of DeblurNVS. 1) We construct a large-scale motion-blurred dataset from DL3DV-10K via frame interpolation and temporal averaging. 2) Given motion-blurred inputs, DeblurNVS learns sharp multi-view latents by jointly adapting the DA3 backbone and multi-view diffusion. 3) Camera tokens are further introduced to guide novel-view latent synthesis, and a shared color decoder maps the restored and synthe… view at source ↗
Figure 4
Figure 4. Figure 4: 2D Deblur+GLD vs Ours. Left: blurred inputs, Uformer-restored inputs, and our restored inputs. Right: ground-truth novel view, Uformer+GLD result, and ours. Independent 2D deblurring remains limited and multi-view inconsistent, leading to degraded GLD synthesis, while our method produces more consistent restorations and sharper novel views. B. Main Results Quantitative Comparisons. We present quantitative … view at source ↗
Figure 5
Figure 5. Figure 5: Qualitative results on the DeblurNeRF-Blender dataset. Compared with baseline methods, our method reconstructs sharper details and achieves higher visual fidelity under motion-blurred inputs. faster than GLD because we simplify the diffusion framework by reducing the number of denoising steps and removing cascaded operations across DA3 feature levels. However, as a diffusion-based method, DeblurNVS is natu… view at source ↗
Figure 6
Figure 6. Figure 6: Qualitative results on the DeblurNeRF-Real dataset. Compared with baseline methods, our method reconstructs sharper details and achieves higher visual fidelity under motion-blurred inputs. Ground Truth Ours w/o DA3LoRA w/o Context Learning [PITH_FULL_IMAGE:figures/full_fig_p008_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: Ablation study on overall architecture. Removing LoRA adaptation from the DA3 backbone or discarding the first-stage context learning leads to degraded geometric structures. TABLE VI COMPARISON WITH THE RGB-SPACE CASCADED BASELINE. UFORMER+GLD FIRST DEBLURS THE INPUT VIEWS WITH UFORMER AND THEN PERFORMS NVS USING GLD. Method PSNR↑ SSIM↑ LPIPS↓ DISTS↓ FID↓ GLD 12.868 0.329 0.570 0.281 166.513 Uformer+GLD 12… view at source ↗
Figure 8
Figure 8. Figure 8: Ablation on hyperparameters of DL3DV-10K-Blur. Columns correspond to different averaging window sizes, while rows correspond to different frame interpolation rates. multi-view deblurring from novel view synthesis. We further construct DL3DV-10K-Blur to support large-scale end-to-end training. Experiments on synthetic and real-world benchmarks demonstrate that DeblurNVS achieves better perceptual quality an… view at source ↗
read the original abstract

Novel view synthesis (NVS) is a fundamental problem in computer vision and graphics. Recent advances in neural radiance fields (NeRF), 3D Gaussian Splatting (3DGS), and generative view synthesis have substantially improved its quality. Yet most methods still rely on clean observations, where image structures and cross-view geometric cues are well preserved. Motion blur breaks this assumption by corrupting local details and weakening multi-view correspondences. Such blur commonly arises from camera shake, scene motion, or finite exposure in practical capture. Blur-aware NVS methods address this degradation by modeling image formation, but their reliance on costly per-scene optimization limits efficient and generalizable sparse-view synthesis. To address this, we propose DeblurNVS, a novel framework for synthesizing high-fidelity novel views directly from sparse motion-blurred images, without requiring per-scene optimization. DeblurNVS restores the intermediate geometric representations needed for multi-view reasoning, enabling blurred inputs to recover reliable structure and correspondence cues. The restored representations are then combined with target camera information to synthesize the target-view representation and reconstruct a sharp RGB novel view. To enable the large-scale training, we construct a motion-blurred NVS dataset from DL3DV-10K using interpolation-based finite-exposure blur synthesis. Extensive experiments demonstrate that DeblurNVS outperforms existing baselines on synthetic motion-blur benchmarks and generalizes to real motion-blurred scenes, producing perceptually sharper and structurally more stable novel views while avoiding costly per-scene optimization. Project page: https://github.com/PKU-YuanGroup/DeblurNVS.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 1 minor

Summary. The paper claims to introduce DeblurNVS, a framework for novel view synthesis from sparse motion-blurred images using geometric latent diffusion to restore intermediate geometric representations from the blurred inputs. These restored representations are then combined with target camera information to synthesize sharp RGB novel views. The approach is enabled by constructing a motion-blurred NVS dataset from DL3DV-10K using interpolation-based finite-exposure blur synthesis. The authors report that DeblurNVS outperforms existing baselines on synthetic motion-blur benchmarks, generalizes to real motion-blurred scenes, and produces perceptually sharper and structurally more stable novel views without requiring costly per-scene optimization.

Significance. If the results hold, the work would be significant for the field of novel view synthesis as it tackles the practical challenge of motion blur in input images, which is common in real-world captures. By avoiding per-scene optimization and using a trained model, it could lead to more efficient and scalable NVS methods applicable to degraded inputs.

major comments (2)
  1. [Abstract] The central claim that the geometric latent diffusion step 'restores the intermediate geometric representations needed for multi-view reasoning, enabling blurred inputs to recover reliable structure and correspondence cues' (abstract) is load-bearing for all downstream synthesis. The interpolation-based finite-exposure blur synthesis used to create the training data from DL3DV-10K may not reproduce real blur statistics such as non-linear trajectories, depth variation, and sensor noise; if the restored latents contain geometric errors, the reported gains in structural stability will not hold.
  2. [Experiments (implied)] No quantitative results, error bars, ablation studies, or metrics on geometric accuracy (e.g., correspondence error or ray consistency) are referenced. Without such evidence, the attribution of improved sharpness and stability to the geometric restoration step rather than other factors cannot be verified.
minor comments (1)
  1. The abstract provides a project page link but does not indicate whether code, models, or the constructed dataset will be released to support reproducibility.

Simulated Author's Rebuttal

2 responses · 0 unresolved

Thank you for the constructive feedback. We address the two major comments point by point below, clarifying the role of the geometric restoration and the available evidence while noting where revisions are warranted.

read point-by-point responses

  1. Referee: [Abstract] The central claim that the geometric latent diffusion step 'restores the intermediate geometric representations needed for multi-view reasoning, enabling blurred inputs to recover reliable structure and correspondence cues' (abstract) is load-bearing for all downstream synthesis. The interpolation-based finite-exposure blur synthesis used to create the training data from DL3DV-10K may not reproduce real blur statistics such as non-linear trajectories, depth variation, and sensor noise; if the restored latents contain geometric errors, the reported gains in structural stability will not hold.

    Authors: We acknowledge that interpolation-based blur synthesis is an approximation and does not fully capture real-world effects such as non-linear trajectories or sensor noise. This approach follows common practice in motion deblurring literature for scalable dataset creation. The manuscript demonstrates generalization to real motion-blurred scenes with perceptually sharper and more stable outputs, which provides indirect support for the utility of the restored representations. We will add a dedicated limitations paragraph on the synthetic blur model and include further qualitative analysis on real data to substantiate the structural stability claims. revision: partial

  2. Referee: [Experiments (implied)] No quantitative results, error bars, ablation studies, or metrics on geometric accuracy (e.g., correspondence error or ray consistency) are referenced. Without such evidence, the attribution of improved sharpness and stability to the geometric restoration step rather than other factors cannot be verified.

    Authors: The manuscript presents quantitative comparisons against baselines on synthetic motion-blur benchmarks using standard image quality metrics, plus qualitative evidence of improved sharpness and stability on both synthetic and real scenes. We agree that direct geometric accuracy metrics and ablations would strengthen attribution to the latent diffusion component. In the revised manuscript we will add error bars, ablation studies isolating the geometric restoration, and new quantitative metrics including correspondence error and ray consistency. revision: yes

Circularity Check

0 steps flagged

No circularity: empirical training pipeline with independent evaluation

full rationale

The paper describes a supervised generative model (geometric latent diffusion) trained on a synthetically blurred dataset constructed via interpolation-based finite-exposure synthesis from DL3DV-10K. Performance claims rest on empirical comparisons against baselines on held-out synthetic and real scenes, not on any derivation that reduces to fitted parameters renamed as predictions or self-referential definitions. No load-bearing step equates the output to the input by construction; the geometric restoration is learned from data rather than presupposed. This is a standard end-to-end trained CV pipeline whose validity is externally falsifiable via benchmark metrics.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Abstract provides no information on free parameters, axioms, or invented entities; all fields left empty.

pith-pipeline@v0.9.1-grok · 5825 in / 945 out tokens · 26945 ms · 2026-06-28T17:26:00.421881+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

83 extracted references · 19 canonical work pages · 8 internal anchors

  1. [1]

    Nerf: Representing scenes as neural radiance fields for view synthesis,

    B. Mildenhall, P. P. Srinivasan, M. Tancik, J. T. Barron, R. Ramamoorthi, and R. Ng, “Nerf: Representing scenes as neural radiance fields for view synthesis,”Communications of the ACM, vol. 65, no. 1, pp. 99–106, 2021

    2021

  2. [2]

    3d gaussian splatting for real-time radiance field rendering,

    B. Kerbl, G. Kopanas, T. Leimk ¨uhler, and G. Drettakis, “3d gaussian splatting for real-time radiance field rendering,”ACM Transactions on Graphics, vol. 42, no. 4, July 2023. [Online]. Available: https://repo-sam.inria.fr/fungraph/3d-gaussian-splatting/

    2023

  3. [3]

    ViewCrafter: Taming Video Diffusion Models for High-fidelity Novel View Synthesis

    W. Yu, J. Xing, L. Yuan, W. Hu, X. Li, Z. Huang, X. Gao, T.-T. Wong, Y . Shan, and Y . Tian, “Viewcrafter: Taming video diffusion models for high-fidelity novel view synthesis,”arXiv preprint arXiv:2409.02048, 2024

    work page internal anchor Pith review Pith/arXiv arXiv 2024

  4. [4]

    Futuregs: Structured gaussian fields for future-aware dynamic scene modeling,

    M. Ding, Z. Wang, J. Wang, T. Han, X. Hu, J. Ding, M. Tan, and Z. Kuang, “Futuregs: Structured gaussian fields for future-aware dynamic scene modeling,” inProceedings of the 33rd ACM International Conference on Multimedia, 2025, pp. 8322–8331

    2025

  5. [5]

    Sparse4dgs: 4d gaussian splatting for sparse-frame dynamic scene reconstruction,

    C. Shi, C. Yang, X. Hu, M. Chen, W. Pan, Y . Yang, J. Ding, Z. Yu, and J. Yu, “Sparse4dgs: 4d gaussian splatting for sparse-frame dynamic scene reconstruction,” inProceedings of the AAAI Conference on Arti- ficial Intelligence, vol. 40, no. 11, 2026, pp. 8933–8941

    2026

  6. [6]

    Drivingforward: Feed-forward 3d gaussian splatting for driving scene reconstruction from flexible surround-view input,

    Q. Tian, X. Tan, Y . Xie, and L. Ma, “Drivingforward: Feed-forward 3d gaussian splatting for driving scene reconstruction from flexible surround-view input,” inProceedings of the AAAI Conference on Ar- tificial Intelligence, vol. 39, no. 7, 2025, pp. 7374–7382

    2025

  7. [7]

    Dggt: Feedforward 4d reconstruction of dynamic driving scenes using unposed images,

    X. Chen, Z. Xiong, Y . Chen, G. Li, N. Wang, H. Luo, L. Chen, H. Sun, B. Wang, G. Chenet al., “Dggt: Feedforward 4d reconstruction of dynamic driving scenes using unposed images,”arXiv preprint arXiv:2512.03004, 2025

    work page arXiv 2025

  8. [8]

    Evolsplat: Efficient volume-based gaussian splatting for urban view synthesis,

    S. Miao, J. Huang, D. Bai, X. Yan, H. Zhou, Y . Wang, B. Liu, A. Geiger, and Y . Liao, “Evolsplat: Efficient volume-based gaussian splatting for urban view synthesis,” inProceedings of the Computer Vision and Pattern Recognition Conference, 2025, pp. 11 286–11 296

    2025

  9. [9]

    Radiance fields in xr: A survey on how radiance fields are envisioned and addressed for xr research,

    K. Li, M. Masuda, S. Schmidt, and S. Mori, “Radiance fields in xr: A survey on how radiance fields are envisioned and addressed for xr research,”IEEE Transactions on Visualization and Computer Graphics, 2025

    2025

  10. [10]

    Huang, S., Chen, L., Zhou, P., Chen, S., Jiang, Z., Hu, Y ., Liao, Y ., Gao, P., Li, H., Yao, M., and Ren, G

    S. Huang, L. Chen, P. Zhou, S. Chen, Z. Jiang, Y . Hu, Y . Liao, P. Gao, H. Li, M. Yaoet al., “Enerverse: Envisioning embodied future space for robotics manipulation,”arXiv preprint arXiv:2501.01895, 2025

    work page arXiv 2025

  11. [11]

    Deepverse: 4d autoregressive video generation as a world model,

    J. Chen, H. Zhu, X. He, Y . Wang, J. Zhou, W. Chang, Y . Zhou, Z. Li, Z. Fu, J. Panget al., “Deepverse: 4d autoregressive video generation as a world model,”arXiv preprint arXiv:2506.01103, 2025

    work page arXiv 2025

  12. [12]

    Aether: Geometric-aware unified world modeling,

    H. Zhu, Y . Wang, J. Zhou, W. Chang, Y . Zhou, Z. Li, J. Chen, C. Shen, J. Pang, and T. He, “Aether: Geometric-aware unified world modeling,” inProceedings of the IEEE/CVF International Conference on Computer Vision, 2025, pp. 8535–8546

    2025

  13. [13]

    Gaussian grouping: Segment and edit anything in 3d scenes,

    M. Ye, M. Danelljan, F. Yu, and L. Ke, “Gaussian grouping: Segment and edit anything in 3d scenes,” inEuropean conference on computer vision. Springer, 2024, pp. 162–179

    2024

  14. [14]

    Gaussianeditor: Editing 3d gaussians delicately with text instructions,

    J. Wang, J. Fang, X. Zhang, L. Xie, and Q. Tian, “Gaussianeditor: Editing 3d gaussians delicately with text instructions,” inProceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2024, pp. 20 902–20 911

    2024

  15. [15]

    Realm: An mllm-agent framework for open world 3d reasoning segmentation and editing on gaussian splatting,

    C. Shi, M. Chen, Y . Mao, C. Yang, X. Hu, J. Ding, and Z. Yu, “Realm: An mllm-agent framework for open world 3d reasoning segmentation and editing on gaussian splatting,”arXiv preprint arXiv:2510.16410, 2025

    work page arXiv 2025

  16. [16]

    Repurposing geometric foundation models for multi-view diffusion,

    W. Jang, S. Jeon, J. Han, J. Choi, M. Kwon, S. Kim, S. Xie, and S. Liu, “Repurposing geometric foundation models for multi-view diffusion,” arXiv preprint arXiv:2603.22275, 2026

    work page arXiv 2026

  17. [17]

    Deblur-nerf: Neural radiance fields from blurry images,

    L. Ma, X. Li, J. Liao, Q. Zhang, X. Wang, J. Wang, and P. V . Sander, “Deblur-nerf: Neural radiance fields from blurry images,” inProceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2022, pp. 12 861–12 870

    2022

  18. [18]

    Bad-nerf: Bundle adjusted deblur neural radiance fields,

    P. Wang, L. Zhao, R. Ma, and P. Liu, “Bad-nerf: Bundle adjusted deblur neural radiance fields,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2023, pp. 4170–4179

    2023

  19. [19]

    Sharp-nerf: Grid-based fast deblurring neural radiance fields using sharpness prior,

    B. Lee, H. Lee, U. Ali, and E. Park, “Sharp-nerf: Grid-based fast deblurring neural radiance fields using sharpness prior,” inProceedings of the IEEE/CVF winter conference on applications of computer vision, 2024, pp. 3709–3718

    2024

  20. [20]

    Bad-gaussians: Bundle adjusted de- blur gaussian splatting,

    L. Zhao, P. Wang, and P. Liu, “Bad-gaussians: Bundle adjusted de- blur gaussian splatting,” inEuropean Conference on Computer Vision. Springer, 2024, pp. 233–250

    2024

  21. [21]

    Bags: Blur agnostic gaussian splatting through multi-scale kernel modeling,

    C. Peng, Y . Tang, Y . Zhou, N. Wang, X. Liu, D. Li, and R. Chellappa, “Bags: Blur agnostic gaussian splatting through multi-scale kernel modeling,” inEuropean Conference on Computer Vision. Springer, 2024, pp. 293–310

    2024

  22. [22]

    Mvsplat: Efficient 3d gaussian splatting from sparse multi-view images,

    Y . Chen, H. Xu, C. Zheng, B. Zhuang, M. Pollefeys, A. Geiger, T.- J. Cham, and J. Cai, “Mvsplat: Efficient 3d gaussian splatting from sparse multi-view images,” inEuropean conference on computer vision. Springer, 2024, pp. 370–386

    2024

  23. [23]

    Anysplat: Feed-forward 3d gaussian splatting from unconstrained views,

    L. Jiang, Y . Mao, L. Xu, T. Lu, K. Ren, Y . Jin, X. Xu, M. Yu, J. Pang, F. Zhaoet al., “Anysplat: Feed-forward 3d gaussian splatting from unconstrained views,”ACM Transactions on Graphics (TOG), vol. 44, no. 6, pp. 1–16, 2025

    2025

  24. [24]

    Mvsplat360: Feed-forward 360 scene synthesis from sparse views,

    Y . Chen, C. Zheng, H. Xu, B. Zhuang, A. Vedaldi, T.-J. Cham, and J. Cai, “Mvsplat360: Feed-forward 360 scene synthesis from sparse views,” in Advances in Neural Information Processing Systems, vol. 37, 2024

    2024

  25. [25]

    Trajectorycrafter: Redirecting camera trajectory for monocular videos via diffusion models,

    M. Yu, W. Hu, J. Xing, and Y . Shan, “Trajectorycrafter: Redirecting camera trajectory for monocular videos via diffusion models,” inPro- ceedings of the IEEE/CVF international conference on computer vision, 2025, pp. 100–111

    2025

  26. [26]

    Deep multi-scale convolutional neural network for dynamic scene deblurring,

    S. Nah, T. Hyun Kim, and K. Mu Lee, “Deep multi-scale convolutional neural network for dynamic scene deblurring,” inProceedings of the IEEE conference on computer vision and pattern recognition, 2017, pp. 3883–3891

    2017

  27. [27]

    Human- aware motion deblurring,

    Z. Shen, W. Wang, X. Lu, J. Shen, H. Ling, T. Xu, and L. Shao, “Human- aware motion deblurring,” inProceedings of the IEEE/CVF International Conference on Computer Vision, 2019, pp. 5572–5581

    2019

  28. [28]

    Real-world blur dataset for learning and benchmarking deblurring algorithms,

    J. Rim, H. Lee, J. Won, and S. Cho, “Real-world blur dataset for learning and benchmarking deblurring algorithms,” inProceedings of the European Conference on Computer Vision, 2020, pp. 184–201

    2020

  29. [29]

    Ntire 2019 challenge on video deblurring and super-resolution: Dataset and study,

    S. Nah, S. Baik, S. Hong, G. Moon, S. Son, R. Timofte, and K. M. Lee, “Ntire 2019 challenge on video deblurring and super-resolution: Dataset and study,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops, 2019

    2019

  30. [30]

    Dl3dv-10k: A large-scale scene dataset for deep learning-based 3d vision,

    L. Ling, Y . Sheng, Z. Tu, W. Zhao, C. Xin, K. Wan, L. Yu, Q. Guo, Z. Yu, Y . Luet al., “Dl3dv-10k: A large-scale scene dataset for deep learning-based 3d vision,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2024, pp. 22 160–22 169

    2024

  31. [31]

    Instant neural graphics primitives with a multiresolution hash encoding,

    T. M ¨uller, A. Evans, C. Schied, and A. Keller, “Instant neural graphics primitives with a multiresolution hash encoding,”ACM transactions on graphics (TOG), vol. 41, no. 4, pp. 1–15, 2022

    2022

  32. [32]

    Tensorf: Tensorial radiance fields,

    A. Chen, Z. Xu, A. Geiger, J. Yu, and H. Su, “Tensorf: Tensorial radiance fields,” inEuropean conference on computer vision. Springer, 2022, pp. 333–350

    2022

  33. [33]

    Mip-nerf: A multiscale representation for anti- aliasing neural radiance fields,

    J. T. Barron, B. Mildenhall, M. Tancik, P. Hedman, R. Martin-Brualla, and P. P. Srinivasan, “Mip-nerf: A multiscale representation for anti- aliasing neural radiance fields,” inProceedings of the IEEE/CVF inter- national conference on computer vision, 2021, pp. 5855–5864

    2021

  34. [34]

    Synsin: End-to-end view synthesis from a single image,

    O. Wiles, G. Gkioxari, R. Szeliski, and J. Johnson, “Synsin: End-to-end view synthesis from a single image,” inProceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2020, pp. 7467– 7477

    2020

  35. [35]

    Pixelsynth: Generating a 3d-consistent experience from a single image,

    C. Rockwell, D. F. Fouhey, and J. Johnson, “Pixelsynth: Generating a 3d-consistent experience from a single image,” inProceedings of the IEEE/CVF International Conference on Computer Vision, 2021, pp. 14 104–14 113

    2021

  36. [36]

    Bridging implicit and explicit geometric transformation for single-image view synthesis,

    B. Park, H. Go, and C. Kim, “Bridging implicit and explicit geometric transformation for single-image view synthesis,”IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 46, no. 9, pp. 6326– 6340, 2024

    2024

  37. [37]

    Stereo Magnification: Learning View Synthesis using Multiplane Images

    T. Zhou, R. Tucker, J. Flynn, G. Fyffe, and N. Snavely, “Stereo magnification: Learning view synthesis using multiplane images,”arXiv preprint arXiv:1805.09817, 2018

    work page internal anchor Pith review Pith/arXiv arXiv 2018

  38. [38]

    Attention is all you need,

    A. Vaswani, N. Shazeer, N. Parmar, J. Uszkoreit, L. Jones, A. N. Gomez, Ł. Kaiser, and I. Polosukhin, “Attention is all you need,”Advances in neural information processing systems, vol. 30, 2017

    2017

  39. [39]

    pixelnerf: Neural radiance fields from one or few images,

    A. Yu, V . Ye, M. Tancik, and A. Kanazawa, “pixelnerf: Neural radiance fields from one or few images,” inProceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2021, pp. 4578– 4587

    2021

  40. [40]

    pixelsplat: 3d gaussian splats from image pairs for scalable generalizable 3d re- construction,

    D. Charatan, S. L. Li, A. Tagliasacchi, and V . Sitzmann, “pixelsplat: 3d gaussian splats from image pairs for scalable generalizable 3d re- construction,” inProceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2024, pp. 19 457–19 467

    2024

  41. [41]

    Vggt: Visual geometry grounded transformer,

    J. Wang, M. Chen, N. Karaev, A. Vedaldi, C. Rupprecht, and D. Novotny, “Vggt: Visual geometry grounded transformer,” inProceedings of the Computer Vision and Pattern Recognition Conference, 2025, pp. 5294– 5306. 11

    2025

  42. [42]

    $\pi^3$: Permutation-Equivariant Visual Geometry Learning

    Y . Wang, J. Zhou, H. Zhu, W. Chang, Y . Zhou, Z. Li, J. Chen, J. Pang, C. Shen, and T. He, “π 3: Permutation-equivariant visual geometry learning,”arXiv preprint arXiv:2507.13347, 2025

    work page internal anchor Pith review Pith/arXiv arXiv 2025

  43. [43]

    Depth anything with any prior,

    Z. Wang, S. Chen, L. Yang, J. Wang, Z. Zhang, H. Zhao, and Z. Zhao, “Depth anything with any prior,”arXiv preprint arXiv:2505.10565, 2025

    work page arXiv 2025

  44. [44]

    FastVGGT: Training-Free Acceleration of Visual Geometry Transformer

    Y . Shen, Z. Zhang, Y . Qu, X. Zheng, J. Ji, S. Zhang, and L. Cao, “Fastvggt: Training-free acceleration of visual geometry transformer,” arXiv preprint arXiv:2509.02560, 2025

    work page internal anchor Pith review Pith/arXiv arXiv 2025

  45. [45]

    Depth Anything 3: Recovering the Visual Space from Any Views

    H. Lin, S. Chen, J. Liew, D. Y . Chen, Z. Li, G. Shi, J. Feng, and B. Kang, “Depth anything 3: Recovering the visual space from any views,”arXiv preprint arXiv:2511.10647, 2025

    work page internal anchor Pith review Pith/arXiv arXiv 2025

  46. [46]

    DINOv2: Learning Robust Visual Features without Supervision

    M. Oquab, T. Darcet, T. Moutakanni, H. V o, M. Szafraniec, V . Khalidov, P. Fernandez, D. Haziza, F. Massa, A. El-Noubyet al., “Dinov2: Learning robust visual features without supervision,”arXiv preprint arXiv:2304.07193, 2023

    work page internal anchor Pith review Pith/arXiv arXiv 2023

  47. [47]

    Denoising diffusion probabilistic models,

    J. Ho, A. Jain, and P. Abbeel, “Denoising diffusion probabilistic models,” Advances in neural information processing systems, vol. 33, pp. 6840– 6851, 2020

    2020

  48. [48]

    Denoising Diffusion Implicit Models

    J. Song, C. Meng, and S. Ermon, “Denoising diffusion implicit models,” arXiv preprint arXiv:2010.02502, 2020

    work page internal anchor Pith review Pith/arXiv arXiv 2010

  49. [49]

    High- resolution image synthesis with latent diffusion models,

    R. Rombach, A. Blattmann, D. Lorenz, P. Esser, and B. Ommer, “High- resolution image synthesis with latent diffusion models,” inProceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2022, pp. 10 684–10 695

    2022

  50. [50]

    Generative novel view synthesis with 3d-aware diffusion models,

    E. R. Chan, K. Nagano, M. A. Chan, A. W. Bergman, J. J. Park, A. Levy, M. Aittala, S. De Mello, T. Karras, and G. Wetzstein, “Generative novel view synthesis with 3d-aware diffusion models,” inProceedings of the IEEE/CVF International Conference on Computer Vision, 2023, pp. 4217–4229

    2023

  51. [51]

    Zero-1-to-3: Zero-shot one image to 3d object,

    R. Liu, R. Wu, B. Van Hoorick, P. Tokmakov, S. Zakharov, and C. V on- drick, “Zero-1-to-3: Zero-shot one image to 3d object,” inProceedings of the IEEE/CVF international conference on computer vision, 2023, pp. 9298–9309

    2023

  52. [52]

    Zeronvs: Zero-shot 360- degree view synthesis from a single image,

    K. Sargent, Z. Li, T. Shah, C. Herrmann, H.-X. Yu, Y . Zhang, E. R. Chan, D. Lagun, L. Fei-Fei, D. Sunet al., “Zeronvs: Zero-shot 360- degree view synthesis from a single image,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2024, pp. 9420–9429

    2024

  53. [53]

    Reconfusion: 3d reconstruction with diffusion priors,

    R. Wu, B. Mildenhall, P. Henzler, K. Park, R. Gao, D. Watson, P. P. Srinivasan, D. Verbin, J. T. Barron, B. Pooleet al., “Reconfusion: 3d reconstruction with diffusion priors,” inProceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2024, pp. 21 551–21 561

    2024

  54. [54]

    Image restoration using very deep convolutional encoder-decoder networks with symmetric skip connec- tions,

    X. Mao, C. Shen, and Y .-B. Yang, “Image restoration using very deep convolutional encoder-decoder networks with symmetric skip connec- tions,”Advances in neural information processing systems, vol. 29, 2016

    2016

  55. [55]

    Beyond a gaussian denoiser: Residual learning of deep cnn for image denoising,

    K. Zhang, W. Zuo, Y . Chen, D. Meng, and L. Zhang, “Beyond a gaussian denoiser: Residual learning of deep cnn for image denoising,”IEEE transactions on image processing, vol. 26, no. 7, pp. 3142–3155, 2017

    2017

  56. [56]

    Enhanced deep residual networks for single image super-resolution,

    B. Lim, S. Son, H. Kim, S. Nah, and K. Mu Lee, “Enhanced deep residual networks for single image super-resolution,” inProceedings of the IEEE conference on computer vision and pattern recognition workshops, 2017, pp. 136–144

    2017

  57. [57]

    NTIRE 2026 3D Restoration and Reconstruction in Real-world Adverse Conditions: RealX3D Challenge Results

    S. Liu, C. Bao, Z. Cui, X. Chu, B. Ren, L. Gu, X. Chen, M. Li, L. Ma, M. V . Condeet al., “Ntire 2026 3d restoration and reconstruction in real-world adverse conditions: Realx3d challenge results,”arXiv preprint arXiv:2604.04135, 2026

    work page internal anchor Pith review Pith/arXiv arXiv 2026

  58. [58]

    R3evision: A survey on robust rendering, restoration, and enhancement for 3d low-level vision,

    W. Kwon, J. Sung, M. Jeon, C. Eom, and J. Oh, “R3evision: A survey on robust rendering, restoration, and enhancement for 3d low-level vision,” arXiv preprint arXiv:2506.16262, 2025

    work page arXiv 2025

  59. [59]

    Breaking the vicious cycle: Coherent 3d gaussian splatting from sparse and motion-blurred views,

    Z. Xu*, C. Feng*, Y . Li, J. Zhao, J. Yang, W. Yu, L. Yuan, and Y . Tian, “Breaking the vicious cycle: Coherent 3d gaussian splatting from sparse and motion-blurred views,”arXiv preprint arXiv:2512.10369, 2025

    work page arXiv 2025

  60. [60]

    Evagaussians: Event stream assisted gaussian splatting from blurry images,

    W. Yu*, C. Feng*, J. Li, J. Tang, J. Yang, Z. Tang, M. Cao, X. Jia, Y . Yang, L. Yuanet al., “Evagaussians: Event stream assisted gaussian splatting from blurry images,” inICCV 2025, 2025

    2025

  61. [61]

    Denoisesplat: Feed-forward gaussian splatting for noisy 3d scene reconstruction,

    F. Jiang, Z. Li, and Y . Zhang, “Denoisesplat: Feed-forward gaussian splatting for noisy 3d scene reconstruction,” inProceedings of the 2026 International Conference on Human-Computer Interaction, Neural Networks and Deep Learning, 2026, pp. 63–71

    2026

  62. [62]

    Srgs: Super-resolution 3d gaussian splatting,

    X. Feng, Y . He, L. Chen, Y . Yang, C. Wang, Y . Chen, Y . Zhong, Z. Kuang, X. Yin, Y . Zhuet al., “Srgs: Super-resolution 3d gaussian splatting,”arXiv preprint arXiv:2404.10318, 2024

    work page arXiv 2024

  63. [63]

    Srsplat: Feed-forward super-resolution gaussian splatting from sparse multi-view images,

    X. Hu, C. Shi, C. Yang, M. Chen, J. Ding, T. Wei, C. Wei, Z. Yu, and M. Tan, “Srsplat: Feed-forward super-resolution gaussian splatting from sparse multi-view images,” inProceedings of the AAAI Conference on Artificial Intelligence, vol. 40, no. 6, 2026, pp. 4950–4958

    2026

  64. [64]

    Sr3r: Rethinking super-resolution 3d reconstruction with feed-forward gaussian splatting,

    X. Feng, X. Wang, T. Zhong, C. Wang, Y . Zhao, T. Xu, Z. Kuang, F. Qin, X. Yin, and Y . Zhu, “Sr3r: Rethinking super-resolution 3d reconstruction with feed-forward gaussian splatting,”arXiv preprint arXiv:2602.24020, 2026

    work page arXiv 2026

  65. [65]

    Learning to extract a video sequence from a single motion-blurred image,

    M. Jin, G. Meishvili, and P. Favaro, “Learning to extract a video sequence from a single motion-blurred image,” inProceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2018, pp. 6334–6342

    2018

  66. [66]

    Scale-recurrent network for deep image deblurring,

    X. Tao, H. Gao, X. Shen, J. Wang, and J. Jia, “Scale-recurrent network for deep image deblurring,” inProceedings of the IEEE conference on computer vision and pattern recognition, 2018, pp. 8174–8182

    2018

  67. [67]

    Deblurgan-v2: De- blurring (orders-of-magnitude) faster and better,

    O. Kupyn, T. Martyniuk, J. Wu, and Z. Wang, “Deblurgan-v2: De- blurring (orders-of-magnitude) faster and better,” inProceedings of the IEEE/CVF international conference on computer vision, 2019, pp. 8878– 8887

    2019

  68. [68]

    Deep video deblurring for hand-held cameras,

    S. Su, M. Delbracio, J. Wang, G. Sapiro, W. Heidrich, and O. Wang, “Deep video deblurring for hand-held cameras,” inProceedings of the IEEE conference on computer vision and pattern recognition, 2017, pp. 1279–1288

    2017

  69. [69]

    Cascaded deep video deblurring using temporal sharpness prior,

    J. Pan, H. Bai, and J. Tang, “Cascaded deep video deblurring using temporal sharpness prior,” inProceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2020, pp. 3043–3051

    2020

  70. [70]

    Dynamic scene deblurring with parameter selective sharing and nested skip connections,

    H. Gao, X. Tao, X. Shen, and J. Jia, “Dynamic scene deblurring with parameter selective sharing and nested skip connections,” inProceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2019, pp. 3848–3856

    2019

  71. [71]

    Restormer: Efficient transformer for high-resolution image restoration,

    S. W. Zamir, A. Arora, S. Khan, M. Hayat, F. S. Khan, and M.-H. Yang, “Restormer: Efficient transformer for high-resolution image restoration,” inProceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2022, pp. 5728–5739

    2022

  72. [72]

    Deblurdiff: Real-word image deblurring with generative diffusion models,

    L. Kong, D. Zou, F. L. Wang, J. Ren, X. Wu, J. Dong, J. Pan et al., “Deblurdiff: Real-word image deblurring with generative diffusion models,” inThe Thirty-ninth Annual Conference on Neural Information Processing Systems, 2025

    2025

  73. [73]

    Mvgenmaster: Scaling multi-view generation from any image via 3d priors enhanced diffusion model,

    C. Cao, C. Yu, S. Liu, F. Wang, X. Xue, and Y . Fu, “Mvgenmaster: Scaling multi-view generation from any image via 3d priors enhanced diffusion model,” inProceedings of the Computer Vision and Pattern Recognition Conference, 2025, pp. 6045–6056

    2025

  74. [74]

    Aligned novel view image and geometry synthesis via cross-modal attention instillation,

    M.-S. Kwak, J. Kim, S. Yun, D. Han, T. Kim, S. Kim, and J.-H. Kim, “Aligned novel view image and geometry synthesis via cross-modal attention instillation,”arXiv preprint arXiv:2506.11924, 2025

    work page arXiv 2025

  75. [75]

    Genwarp: Single image to novel views with semantic-preserving generative warping,

    J. Seo, K. Fukuda, T. Shibuya, T. Narihira, N. Murata, S. Hu, C.- H. Lai, S. Kim, and Y . Mitsufuji, “Genwarp: Single image to novel views with semantic-preserving generative warping,”Advances in Neural Information Processing Systems, vol. 37, pp. 80 220–80 243, 2024

    2024

  76. [76]

    Revisiting adaptive convolutions for video frame interpolation,

    S. Niklaus, L. Mai, and O. Wang, “Revisiting adaptive convolutions for video frame interpolation,” inProceedings of the IEEE/CVF winter conference on applications of computer vision, 2021, pp. 1099–1109

    2021

  77. [77]

    Lora: Low-rank adaptation of large language models

    E. J. Hu, Y . Shen, P. Wallis, Z. Allen-Zhu, Y . Li, S. Wang, L. Wang, W. Chenet al., “Lora: Low-rank adaptation of large language models.” Iclr, vol. 1, no. 2, p. 3, 2022

    2022

  78. [78]

    The unreasonable effectiveness of deep features as a perceptual metric,

    R. Zhang, P. Isola, A. A. Efros, E. Shechtman, and O. Wang, “The unreasonable effectiveness of deep features as a perceptual metric,” in Proceedings of the IEEE conference on computer vision and pattern recognition, 2018, pp. 586–595

    2018

  79. [79]

    Generative adversarial nets,

    I. J. Goodfellow, J. Pouget-Abadie, M. Mirza, B. Xu, D. Warde-Farley, S. Ozair, A. Courville, and Y . Bengio, “Generative adversarial nets,” Advances in neural information processing systems, vol. 27, 2014

    2014

  80. [80]

    Image quality assessment: From error visibility to structural similarity,

    Z. Wang, A. C. Bovik, H. R. Sheikh, and E. P. Simoncelli, “Image quality assessment: From error visibility to structural similarity,”IEEE Transactions on Image Processing, vol. 13, no. 4, pp. 600–612, 2004

    2004

Showing first 80 references.