pith. machine review for the scientific record. sign in

arxiv: 2604.08500 · v1 · submitted 2026-04-09 · 💻 cs.CV

Recognition: no theorem link

Novel View Synthesis as Video Completion

Qi Wu , Khiem Vuong , Minsik Jeon , Srinivasa Narasimhan , Deva Ramanan
Authors on Pith no claims yet

Pith reviewed 2026-05-10 18:21 UTC · model grok-4.3

classification 💻 cs.CV
keywords novel view synthesisvideo diffusion modelsvideo completionpermutation invariancesparse multi-viewtemporal embeddingsframe ordering
0
0 comments X

The pith

Video diffusion models can be adapted for sparse novel view synthesis by reformulating it as low frame-rate video completion and removing temporal order cues.

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

The paper sets out to show that existing video diffusion models already hold implicit knowledge of multi-view geometry, which can be activated for novel view synthesis without building new image-based priors from scratch. It does this by casting the task of predicting a target view from a few unordered input images and their poses as a kind of sparse video completion problem. The key adaptation is to make the model ignore the usual frame ordering so it treats the inputs as a permutation-invariant set rather than a timed sequence. If this holds, it means practitioners can repurpose large video models for 3D view synthesis with only light fine-tuning instead of training dedicated multi-view generators.

Core claim

Sparse novel view synthesis reduces to low frame-rate video completion once video diffusion models are modified to become invariant to input ordering; the modifications consist of per-frame latent encodings plus removal of temporal positional embeddings, after which the models can be fine-tuned with minimal supervision to produce competitive results on standard sparse-view benchmarks.

What carries the argument

FrameCrafter, the adapted video diffusion architecture that uses per-frame latent encodings and drops temporal positional embeddings to enforce permutation invariance on unordered input sets.

If this is right

  • Video models contain implicit multi-view geometric knowledge that survives the removal of temporal positional embeddings.
  • Only minimal supervision is needed to train such models to ignore frame ordering while still producing coherent novel views.
  • Sparse-view NVS performance becomes competitive with methods that rely on single-image generative priors.
  • The same adaptation pattern may allow other video-trained models to handle unordered multi-view tasks without explicit 3D supervision.

Where Pith is reading between the lines

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

  • The same removal of temporal embeddings could be tested on other tasks that require set-like rather than sequential inputs, such as unordered image sets for 3D reconstruction.
  • If the approach scales, it suggests that future multi-view generators might be derived from video pre-training rather than image pre-training, changing data collection priorities.
  • One could measure how much geometric consistency is retained purely by the diffusion prior versus how much is added during the light fine-tuning stage.

Load-bearing premise

Video diffusion models already encode usable multi-view geometric knowledge from their original training that survives the removal of temporal cues and can be activated by light fine-tuning on sparse inputs.

What would settle it

Train a video diffusion model on the described modifications and fine-tuning regime, then check whether the generated target views remain geometrically consistent with the provided camera poses when the input images are presented in random order; failure of consistency across random permutations would falsify the claim.

Figures

Figures reproduced from arXiv: 2604.08500 by Deva Ramanan, Khiem Vuong, Minsik Jeon, Qi Wu, Srinivasa Narasimhan.

Figure 1
Figure 1. Figure 1: Video Diffusion as Multi-view Prior. While prior generative NVS meth￾ods typically initialize from image diffusion models and rely on large pose-annotated multi-view datasets, videos naturally capture viewpoint changes and cross-view con￾sistency, making them a more scalable source of supervision for NVS. We present FrameCrafter, which adapts pretrained video diffusion models with lightweight mod￾ification… view at source ↗
Figure 2
Figure 2. Figure 2: Each input image is encoded by a frozen video VAE, and its camera pose is converted into a Plücker ray map that is concatenated with the image latent along the channel dimension. The resulting view tokens are then concatenated along the temporal dimension, where the first K tokens correspond to input views and the final token represents the query view. Only the patch embedding layer and the LoRA modules in… view at source ↗
Figure 3
Figure 3. Figure 3: Comparison between standard causal video VAE encoding and our [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Qualitative comparison under sparse inputs. [PITH_FULL_IMAGE:figures/full_fig_p011_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Performance on different num￾bers of training scenes. Even with only 20 scenes, our model surpasses EscherNet trained on 10K scenes. Performance im￾proves consistently as training data in￾creases, demonstrating strong data effi￾ciency and scalability. Data Scaling We further analyze the effect of training data scale on model performance ( [PITH_FULL_IMAGE:figures/full_fig_p013_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Additional Qualitative Comparisons with SEVA [55], Aether [57] and LVSM [17]. serve as a form of data augmentation: each scene can be observed under many differ￾ent input orderings, exposing the model to a much richer set of training configurations. From this perspective, permutation-invariant training can be viewed as implicitly lever￾aging an exponentially larger set of permuted training samples. Our arc… view at source ↗
read the original abstract

We tackle the problem of sparse novel view synthesis (NVS) using video diffusion models; given $K$ ($\approx 5$) multi-view images of a scene and their camera poses, we predict the view from a target camera pose. Many prior approaches leverage generative image priors encoded via diffusion models. However, models trained on single images lack multi-view knowledge. We instead argue that video models already contain implicit multi-view knowledge and so should be easier to adapt for NVS. Our key insight is to formulate sparse NVS as a low frame-rate video completion task. However, one challenge is that sparse NVS is defined over an unordered set of inputs, often too sparse to admit a meaningful order, so the models should be $\textit{invariant}$ to permutations of that input set. To this end, we present FrameCrafter, which adapts video models (naturally trained with coherent frame orderings) to permutation-invariant NVS through several architectural modifications, including per-frame latent encodings and removal of temporal positional embeddings. Our results suggest that video models can be easily trained to "forget" about time with minimal supervision, producing competitive performance on sparse-view NVS benchmarks. Project page: https://frame-crafter.github.io/

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 / 2 minor

Summary. The manuscript proposes FrameCrafter, an adaptation of video diffusion models for sparse novel view synthesis (NVS). It reformulates the task as low frame-rate video completion given K≈5 unordered input views and their camera poses, with the goal of predicting a target view. Architectural changes include per-frame latent encodings and removal of temporal positional embeddings to achieve permutation invariance, allowing the model to 'forget' about time. The central claim is that video models already encode implicit multi-view knowledge and can be adapted with minimal supervision to achieve competitive results on sparse-view NVS benchmarks.

Significance. If the empirical claims hold, the work is significant because it offers a new perspective on leveraging video generative priors for multi-view geometry tasks rather than relying solely on single-image diffusion models. The reformulation as video completion and the focus on minimal supervision are strengths. Credit is due for the explicit handling of unordered inputs via the described modifications, which could influence future work on adapting temporal models to set-based tasks in computer vision.

major comments (2)
  1. [Method (description of per-frame encodings and temporal embedding removal)] The central claim of permutation invariance (and thus successful 'forgetting' of time) rests on the modifications in the FrameCrafter architecture. However, video diffusion backbones typically retain 3D convolutions, temporal attention layers, or frame-wise processing that can retain residual order dependence even after temporal positional embeddings are removed. No experiment is reported that feeds the same K views in different orders and verifies identical outputs. This directly affects whether the competitive benchmark performance can be attributed to implicit multi-view knowledge rather than order artifacts.
  2. [Abstract and Experiments section] The abstract asserts 'competitive performance on sparse-view NVS benchmarks' but the provided text contains no quantitative numbers, specific baselines, ablation tables, or error metrics. Without these details (presumably in §4), the strength of the empirical support for the central claim cannot be assessed. Please add explicit comparisons to prior NVS methods and ablations on the invariance modifications.
minor comments (2)
  1. [Method] Clarify how the target camera pose is encoded and injected into the model, as this is central to the NVS formulation but not detailed in the abstract.
  2. [Experiments] Ensure all figures showing qualitative results include the input views, target pose, and ground truth for direct comparison.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for their constructive and insightful comments. We address each major comment point by point below and describe the revisions we will make to strengthen the manuscript.

read point-by-point responses

  1. Referee: [Method (description of per-frame encodings and temporal embedding removal)] The central claim of permutation invariance (and thus successful 'forgetting' of time) rests on the modifications in the FrameCrafter architecture. However, video diffusion backbones typically retain 3D convolutions, temporal attention layers, or frame-wise processing that can retain residual order dependence even after temporal positional embeddings are removed. No experiment is reported that feeds the same K views in different orders and verifies identical outputs. This directly affects whether the competitive benchmark performance can be attributed to implicit multi-view knowledge rather than order artifacts.

    Authors: We agree that an explicit empirical verification of permutation invariance is valuable and would strengthen the central claim. While the per-frame latent encodings and removal of temporal positional embeddings are specifically designed to eliminate order cues (allowing the model to process inputs as an unordered set), we acknowledge that residual dependencies could potentially remain in other backbone components such as 3D convolutions or attention layers. In the revised manuscript, we will add a dedicated experiment that feeds identical sets of K input views in multiple random permutations and quantifies output consistency (e.g., via PSNR, SSIM, and LPIPS between the resulting target views). This will provide direct evidence that performance arises from implicit multi-view knowledge rather than order artifacts. revision: yes

  2. Referee: [Abstract and Experiments section] The abstract asserts 'competitive performance on sparse-view NVS benchmarks' but the provided text contains no quantitative numbers, specific baselines, ablation tables, or error metrics. Without these details (presumably in §4), the strength of the empirical support for the central claim cannot be assessed. Please add explicit comparisons to prior NVS methods and ablations on the invariance modifications.

    Authors: The full manuscript contains quantitative results and comparisons in Section 4, including evaluations on standard sparse-view NVS benchmarks (DTU and RealEstate10K), direct comparisons against prior methods (e.g., single-image diffusion baselines and dedicated NVS approaches), ablation studies isolating the effects of the per-frame encoding and temporal embedding removal, and standard error metrics (PSNR, SSIM, LPIPS). The abstract summarizes these findings at a high level without numbers, which is conventional for brevity. To make the empirical support more immediately accessible, we will revise the abstract to include a brief mention of key quantitative highlights and ensure the experiments section features clearly labeled tables for all baselines and ablations. revision: yes

Circularity Check

0 steps flagged

Empirical adaptation of video models for NVS is self-contained with no circular derivation

full rationale

The paper proposes treating sparse NVS as low frame-rate video completion and adapts video diffusion models via per-frame latent encodings plus removal of temporal positional embeddings to enforce permutation invariance. Effectiveness is measured by performance on external benchmarks rather than by any internal loss or fitted quantity. No equations, predictions, or uniqueness claims reduce to the method's own inputs by construction. The approach is presented as a practical architectural modification whose validity is established empirically, not definitionally or via self-citation chains.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The central claim rests on the domain assumption that video models already contain usable multi-view knowledge; no new physical entities or free parameters are introduced in the abstract, but training hyperparameters and the exact definition of 'minimal supervision' remain unspecified.

axioms (1)
  • domain assumption Video diffusion models trained on coherent frame sequences already encode implicit multi-view geometric knowledge.
    Explicitly stated as the key insight in the abstract.

pith-pipeline@v0.9.0 · 5526 in / 1274 out tokens · 27309 ms · 2026-05-10T18:21:54.405209+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

57 extracted references · 23 canonical work pages · 10 internal anchors

  1. [1]

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

    Bai,J., Xia, M., Fu, X.,Wang, X.,Mu, L., Cao, J.,Liu, Z., Hu, H., Bai, X., Wan, P., et al.: Recammaster: Camera-controlled generative rendering from a single video. In: Proceedings of the IEEE/CVF International Conference on Computer Vision. pp. 14834–14844 (2025)

    2025

  2. [2]

    In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition

    Barron, J.T., Mildenhall, B., Verbin, D., Srinivasan, P.P., Hedman, P.: Mip- nerf 360: Unbounded anti-aliased neural radiance fields. In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition. pp. 5470–5479 (2022)

    2022

  3. [3]

    Stable Video Diffusion: Scaling Latent Video Diffusion Models to Large Datasets

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

    work page internal anchor Pith review arXiv 2023

  4. [4]

    Brooks, T., Peebles, B., Holmes, C., DePue, W., Guo, Y., Jing, L., Schnurr, D., Taylor, J., Luhman, T., Luhman, E., et al.: Video generation models as world simulators. 2024. URL https://openai. com/research/video-generation-models-as- world-simulators3(1), 3 (2024)

    2024

  5. [5]

    In: Proceedings of the IEEE/CVF international conference on computer vision

    Chen, A., Xu, Z., Zhao, F., Zhang, X., Xiang, F., Yu, J., Su, H.: Mvsnerf: Fast generalizable radiance field reconstruction from multi-view stereo. In: Proceedings of the IEEE/CVF international conference on computer vision. pp. 14124–14133 (2021)

    2021

  6. [6]

    VideoCrafter1: Open Diffusion Models for High-Quality Video Generation

    Chen, H., Xia, M., He, Y., Zhang, Y., Cun, X., Yang, S., Xing, J., Liu, Y., Chen, Q., Wang, X., et al.: Videocrafter1: Open diffusion models for high-quality video generation. arXiv preprint arXiv:2310.19512 (2023)

    work page internal anchor Pith review arXiv 2023

  7. [7]

    arXiv preprint arXiv:2507.12646 (2025)

    Chen, K., Khurana, T., Ramanan, D.: Reconstruct, inpaint, test-time fine- tune: Dynamic novel-view synthesis from monocular videos. arXiv preprint arXiv:2507.12646 (2025)

    work page arXiv 2025

  8. [8]

    Advances in Neural Information Processing Systems36, 35799–35813 (2023)

    Deitke, M., Liu, R., Wallingford, M., Ngo, H., Michel, O., Kusupati, A., Fan, A., Laforte, C., Voleti, V., Gadre, S.Y., et al.: Objaverse-xl: A universe of 10m+ 3d objects. Advances in Neural Information Processing Systems36, 35799–35813 (2023)

    2023

  9. [9]

    Dream- sim: Learning new dimensions of human visual similar- ity using synthetic data.arXiv preprint arXiv:2306.09344,

    Fu,S.,Tamir,N.,Sundaram,S.,Chai,L.,Zhang,R.,Dekel,T.,Isola,P.:Dreamsim: Learning new dimensions of human visual similarity using synthetic data. arXiv preprint arXiv:2306.09344 (2023) 16 Q. Wu et al

    work page arXiv 2023

  10. [10]

    Cat3d: Create any- thing in 3d with multi-view diffusion models,

    Gao, R., Holynski, A., Henzler, P., Brussee, A., Martin-Brualla, R., Srinivasan, P., Barron, J.T., Poole, B.: Cat3d: Create anything in 3d with multi-view diffusion models. arXiv preprint arXiv:2405.10314 (2024)

    work page arXiv 2024

  11. [11]

    He, H., Xu, Y., Guo, Y., Wetzstein, G., Dai, B., Li, H., Yang, C.: Cameractrl: En- ablingcameracontrolfortext-to-videogeneration.arXivpreprintarXiv:2404.02101 (2024)

    work page internal anchor Pith review arXiv 2024

  12. [12]

    Advances in neural information processing systems33, 6840–6851 (2020)

    Ho, J., Jain, A., Abbeel, P.: Denoising diffusion probabilistic models. Advances in neural information processing systems33, 6840–6851 (2020)

    2020

  13. [13]

    Iclr1(2), 3 (2022)

    Hu, E.J., Shen, Y., Wallis, P., Allen-Zhu, Z., Li, Y., Wang, S., Wang, L., Chen, W., et al.: Lora: Low-rank adaptation of large language models. Iclr1(2), 3 (2022)

    2022

  14. [14]

    Huang, Z., Li, X., Lv, Z., Rehg, J.M.: How much 3d do video foundation models encode? arXiv preprint arXiv:2512.19949 (2025)

    work page arXiv 2025

  15. [15]

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

    Jiang, H., Tan, H., Wang, P., Jin, H., Zhao, Y., Bi, S., Zhang, K., Luan, F., Sunkavalli, K., Huang, Q., et al.: Rayzer: A self-supervised large view synthesis model. In: Proceedings of the IEEE/CVF International Conference on Computer Vision. pp. 4918–4929 (2025)

    2025

  16. [16]

    Jiang, Z., Zheng, C., Laina, I., Larlus, D., Vedaldi, A.: Geo4d: Leveraging video generatorsforgeometric4dscenereconstruction.In:ProceedingsoftheIEEE/CVF International Conference on Computer Vision. pp. 20658–20671 (2025)

    2025

  17. [17]

    Lvsm: A large view synthesis model with minimal 3d inductive bias.arXiv preprint arXiv:2410.17242, 2024

    Jin, H., Jiang, H., Tan, H., Zhang, K., Bi, S., Zhang, T., Luan, F., Snavely, N., Xu, Z.: Lvsm: A large view synthesis model with minimal 3d inductive bias. arXiv preprint arXiv:2410.17242 (2024)

    work page arXiv 2024

  18. [18]

    ACM Trans

    Kerbl, B., Kopanas, G., Leimkühler, T., Drettakis, G., et al.: 3d gaussian splatting for real-time radiance field rendering. ACM Trans. Graph.42(4), 139–1 (2023)

    2023

  19. [19]

    HunyuanVideo: A Systematic Framework For Large Video Generative Models

    Kong, W., Tian, Q., Zhang, Z., Min, R., Dai, Z., Zhou, J., Xiong, J., Li, X., Wu, B., Zhang, J., et al.: Hunyuanvideo: A systematic framework for large video generative models. arXiv preprint arXiv:2412.03603 (2024)

    work page internal anchor Pith review Pith/arXiv arXiv 2024

  20. [20]

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

    Kong,X.,Liu,S.,Lyu,X.,Taher,M.,Qi,X.,Davison,A.J.:Eschernet:Agenerative model for scalable view synthesis. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. pp. 9503–9513 (2024)

    2024

  21. [21]

    Cameras as relative positional encoding.arXiv preprint arXiv:2507.10496,

    Li, R., Yi, B., Liu, J., Gao, H., Ma, Y., Kanazawa, A.: Cameras as relative posi- tional encoding. arXiv preprint arXiv:2507.10496 (2025)

    work page arXiv 2025

  22. [22]

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

    Ling, L., Sheng, Y., Tu, Z., Zhao, W., Xin, C., Wan, K., Yu, L., Guo, Q., Yu, Z., Lu, Y., et al.: Dl3dv-10k: A large-scale scene dataset for deep learning-based 3d vision. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. pp. 22160–22169 (2024)

    2024

  23. [23]

    Flow Matching for Generative Modeling

    Lipman, Y., Chen, R.T., Ben-Hamu, H., Nickel, M., Le, M.: Flow matching for generative modeling. arXiv preprint arXiv:2210.02747 (2022)

    work page internal anchor Pith review Pith/arXiv arXiv 2022

  24. [24]

    In: Proceedings of the IEEE/CVF inter- national conference on computer vision

    Liu, R., Wu, R., Van Hoorick, B., Tokmakov, P., Zakharov, S., Vondrick, C.: Zero- 1-to-3: Zero-shot one image to 3d object. In: Proceedings of the IEEE/CVF inter- national conference on computer vision. pp. 9298–9309 (2023)

    2023

  25. [25]

    Syncdreamer: Gen- erating multiview-consistent images from a single-view im- age.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)

    work page arXiv 2023

  26. [26]

    Commu- nications of the ACM65(1), 99–106 (2021)

    Mildenhall, B., Srinivasan, P.P., Tancik, M., Barron, J.T., Ramamoorthi, R., Ng, R.: Nerf: Representing scenes as neural radiance fields for view synthesis. Commu- nications of the ACM65(1), 99–106 (2021)

    2021

  27. [27]

    In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition

    Niemeyer, M., Barron, J.T., Mildenhall, B., Sajjadi, M.S., Geiger, A., Radwan, N.: Regnerf: Regularizing neural radiance fields for view synthesis from sparse in- puts. In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition. pp. 5480–5490 (2022) FrameCrafter 17

    2022

  28. [28]

    Peebles,W.,Xie,S.:Scalablediffusionmodelswithtransformers.In:Proceedingsof the IEEE/CVF international conference on computer vision. pp. 4195–4205 (2023)

    2023

  29. [29]

    on a new geometry of space

    Plucker, J.: Xvii. on a new geometry of space. Philosophical Transactions of the Royal Society of London (155), 725–791 (1865)

  30. [30]

    In: Proceedings of the IEEE/CVF international conference on computer vision

    Reizenstein, J., Shapovalov, R., Henzler, P., Sbordone, L., Labatut, P., Novotny, D.: Common objects in 3d: Large-scale learning and evaluation of real-life 3d cat- egory reconstruction. In: Proceedings of the IEEE/CVF international conference on computer vision. pp. 10901–10911 (2021)

    2021

  31. [31]

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

    Ren,X.,Shen,T.,Huang,J.,Ling,H.,Lu,Y.,Nimier-David,M.,Müller,T.,Keller, A., Fidler, S., Gao, J.: Gen3c: 3d-informed world-consistent video generation with precise camera control. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. pp. 6121–6132 (2025)

    2025

  32. [32]

    In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition

    Rombach, R., Blattmann, A., Lorenz, D., Esser, P., Ommer, B.: High-resolution image synthesis with latent diffusion models. In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition. pp. 10684–10695 (2022)

    2022

  33. [33]

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

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

    2024

  34. [34]

    Zero123++: a single image to consistent multi-view dif- fusion base model.arXiv preprint arXiv:2310.15110, 2023

    Shi, R., Chen, H., Zhang, Z., Liu, M., Xu, C., Wei, X., Chen, L., Zeng, C., Su, H.: Zero123++: a single image to consistent multi-view diffusion base model. arXiv preprint arXiv:2310.15110 (2023)

    work page arXiv 2023

  35. [35]

    In: Proceedings of the IEEE conference on computer vision and pattern recognition

    Shi, W., Caballero, J., Huszár, F., Totz, J., Aitken, A.P., Bishop, R., Rueckert, D., Wang, Z.: Real-time single image and video super-resolution using an efficient sub-pixel convolutional neural network. In: Proceedings of the IEEE conference on computer vision and pattern recognition. pp. 1874–1883 (2016)

    2016

  36. [36]

    Score-Based Generative Modeling through Stochastic Differential Equations

    Song, Y., Sohl-Dickstein, J., Kingma, D.P., Kumar, A., Ermon, S., Poole, B.: Score- based generative modeling through stochastic differential equations. arXiv preprint arXiv:2011.13456 (2020)

    work page internal anchor Pith review Pith/arXiv arXiv 2011

  37. [37]

    Neurocomputing568, 127063 (2024)

    Su, J., Ahmed, M., Lu, Y., Pan, S., Bo, W., Liu, Y.: Roformer: Enhanced trans- former with rotary position embedding. Neurocomputing568, 127063 (2024)

    2024

  38. [38]

    arXiv preprint arXiv:2601.16982 (2026)

    Van Hoorick, B., Chen, D., Iwase, S., Tokmakov, P., Irshad, M.Z., Vasiljevic, I., Gupta, S., Cheng, F., Zakharov, S., Guizilini, V.C.: Anyview: Synthesizing any novel view in dynamic scenes. arXiv preprint arXiv:2601.16982 (2026)

    work page arXiv 2026

  39. [39]

    Wan: Open and Advanced Large-Scale Video Generative Models

    Wan, T., Wang, A., Ai, B., Wen, B., Mao, C., Xie, C.W., Chen, D., Yu, F., Zhao, H., Yang, J., Zeng, J., Wang, J., Zhang, J., Zhou, J., Wang, J., Chen, J., Zhu, K., Zhao, K., Yan, K., Huang, L., Feng, M., Zhang, N., Li, P., Wu, P., Chu, R., Feng, R., Zhang, S., Sun, S., Fang, T., Wang, T., Gui, T., Weng, T., Shen, T., Lin, W., Wang, W., Wang, W., Zhou, W.,...

    work page internal anchor Pith review Pith/arXiv arXiv 2025

  40. [40]

    In: Proceedings of the Computer Vision and Pattern Recognition Conference

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

    2025

  41. [41]

    In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition

    Wang, Q., Wang, Z., Genova, K., Srinivasan, P.P., Zhou, H., Barron, J.T., Martin- Brualla, R., Snavely, N., Funkhouser, T.: Ibrnet: Learning multi-view image-based rendering. In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition. pp. 4690–4699 (2021) 18 Q. Wu et al

    2021

  42. [42]

    In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition

    Wang, S., Leroy, V., Cabon, Y., Chidlovskii, B., Revaud, J.: Dust3r: Geometric 3d vision made easy. In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition. pp. 20697–20709 (2024)

    2024

  43. [43]

    In: ACM SIGGRAPH 2024 Conference Papers

    Wang, Z., Yuan, Z., Wang, X., Li, Y., Chen, T., Xia, M., Luo, P., Shan, Y.: Mo- tionctrl: A unified and flexible motion controller for video generation. In: ACM SIGGRAPH 2024 Conference Papers. pp. 1–11 (2024)

    2024

  44. [44]

    In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition

    Wu, R., Mildenhall, B., Henzler, P., Park, K., Gao, R., Watson, D., Srinivasan, P.P., Verbin, D., Barron, J.T., Poole, B., et al.: Reconfusion: 3d reconstruction with diffusion priors. In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition. pp. 21551–21561 (2024)

    2024

  45. [45]

    arXiv preprint arXiv:2512.03040 (2025)

    Xiao, Z., Zhao, Y., Li, L., Lan, Y., Yu, N., Garg, R., Cooper, R., Taghavi, M.H., Pan,X.:Video4spatial:Towardsvisuospatialintelligencewithcontext-guidedvideo generation. arXiv preprint arXiv:2512.03040 (2025)

    work page arXiv 2025

  46. [46]

    CogVideoX: Text-to-Video Diffusion Models with An Expert Transformer

    Yang, Z., Teng, J., Zheng, W., Ding, M., Huang, S., Xu, J., Yang, Y., Hong, W., Zhang, X., Feng, G., et al.: Cogvideox: Text-to-video diffusion models with an expert transformer. arXiv preprint arXiv:2408.06072 (2024)

    work page internal anchor Pith review Pith/arXiv arXiv 2024

  47. [47]

    In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition

    Yu,A.,Ye,V.,Tancik,M.,Kanazawa, A.:pixelnerf:Neuralradiancefieldsfromone or few images. In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition. pp. 4578–4587 (2021)

    2021

  48. [48]

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

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

    work page internal anchor Pith review arXiv 2024

  49. [49]

    Zhang, Amy Lin, Moneish Kumar, Tzu-Hsuan Yang, Deva Ramanan, and Shubham Tulsiani

    Zhang, J.Y., Lin, A., Kumar, M., Yang, T.H., Ramanan, D., Tulsiani, S.: Cameras as rays: Pose estimation via ray diffusion. arXiv preprint arXiv:2402.14817 (2024)

    work page arXiv 2024

  50. [50]

    In: European Conference on Computer Vision

    Zhang, K., Bi, S., Tan, H., Xiangli, Y., Zhao, N., Sunkavalli, K., Xu, Z.: Gs-lrm: Large reconstruction model for 3d gaussian splatting. In: European Conference on Computer Vision. pp. 1–19. Springer (2024)

    2024

  51. [51]

    In: Proceedings of the IEEE conference on computer vision and pattern recognition

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

    2018

  52. [52]

    arXiv preprint arXiv:2512.10950 (2025) 3

    Zhao, Q., Tan, H., Wang, Q., Bi, S., Zhang, K., Sunkavalli, K., Tulsiani, S., Jiang, H.: E-rayzer: Self-supervised 3d reconstruction as spatial visual pre-training. arXiv preprint arXiv:2512.10950 (2025)

    work page arXiv 2025

  53. [53]

    arXiv preprint arXiv:2410.15957 , year=

    Zheng, G., Li, T., Jiang, R., Lu, Y., Wu, T., Li, X.: Cami2v: Camera-controlled image-to-video diffusion model. arXiv preprint arXiv:2410.15957 (2024)

    work page arXiv 2024

  54. [54]

    In: Proceedings of the SIGGRAPH Asia 2025 Conference Papers

    Zhi, Y., Li, C., Liao, H., Yang, X., Sun, Z., Chang, J., Cun, X., Feng, W., Han, X.: Mv-performer: Taming video diffusion model for faithful and synchronized multi- view performer synthesis. In: Proceedings of the SIGGRAPH Asia 2025 Conference Papers. pp. 1–14 (2025)

    2025

  55. [55]

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

    Zhou,J.,Gao,H.,Voleti,V.,Vasishta,A.,Yao,C.H.,Boss,M.,Torr,P.,Rupprecht, C., Jampani, V.: Stable virtual camera: Generative view synthesis with diffusion models. In: Proceedings of the IEEE/CVF International Conference on Computer Vision. pp. 12405–12414 (2025)

    2025

  56. [56]

    Zhou, T., Tucker, R., Flynn, J., Fyffe, G., Snavely, N.: Stereo magnification: Learn- ingviewsynthesisusingmultiplaneimages.arXivpreprintarXiv:1805.09817(2018)

    work page internal anchor Pith review arXiv 2018

  57. [57]

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

    Zhu, H., Wang, Y., Zhou, J., Chang, W., Zhou, Y., Li, Z., Chen, J., Shen, C., Pang, J., He, T.: Aether: Geometric-aware unified world modeling. In: Proceedings of the IEEE/CVF International Conference on Computer Vision. pp. 8535–8546 (2025) FrameCrafter 19 Appendix Overview.In the appendix, we provide additional quantitative/qualitative results, ablation...

    2025