Pith. sign in

REVIEW 4 major objections 5 minor 1 cited by

4D Gaussian Splatting in the Wild with Uncertainty-Aware Regularization

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 Uncertainty-aware regularization lets 4D Gaussian Splatting handle casually filmed dynamic scenes without overfitting training frames.

desk verdict Useful 4DGS regularization paper with real empirical gains, but the uncertainty map's blind spot at disoccluded regions and the overstated per-scene consistency make the core mechanism weaker than the headline. read the letter →

arxiv 2411.08879 v1 pith:QHIVHONG submitted 2024-11-13 cs.CV cs.AI

Mijeong Kim , Jongwoo Lim , Bohyung Han This is my paper · ORCID
classification cs.CVcs.AI
keywords 4DGaussianSplattingnovelviewsynthesisdynamicscenereconstructionuncertainty-awareregularizationdiffusionpriorsdepthsmoothnessmonocularvideoflowdensification
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

This paper tackles a gap in 4D Gaussian Splatting: reconstructing dynamic scenes from casually recorded monocular videos, where a handheld camera moves around fast-moving objects. Existing 4DGS methods overfit training frames because regularization is applied uniformly, sacrificing reconstruction quality, and SfM initialization leaves fast-moving regions empty. The paper proposes to measure each Gaussian primitive's uncertainty from how much it contributed to training-image rendering, splat that uncertainty into a per-pixel map for unseen views, and use the map to weight a diffusion prior and a depth-smoothness prior. It also seeds additional primitives in dynamic regions using scene flow and estimated depth. The result is a 4DGS pipeline that improves both novel view synthesis and training reconstruction on the DyCheck dataset, and the same uncertainty gating improves few-shot static reconstruction on LLFF.

What carries the argument

The load-bearing object is the contribution-based uncertainty map $U(r)=\sum_k \omega_k^\pi(r)\,U_k$, where each primitive's uncertainty is $U_k=1-\mathrm{Sigmoid}(C_k; c_0,c_1)$ and $C_k=\sum_{I\in\mathcal{T}}\sum_r \omega_k^\pi(r)$ is the primitive's total $\alpha$-blending contribution across training pixels. The same $\alpha$-blending weights that render color also render uncertainty and depth, so the gate is exactly as view-dependent as the image itself; the map then multiplies the diffusion loss and the total-variation depth-smoothing loss, concentrating both on regions the training data constrain least. A second mechanism is dynamic-region densification, which seeds additional Gaussians at scene-flow-detected dynamic pixels using estimated depth, repairing SfM's blind spots.

What would settle it

A decisive test would be to render a synthetic dynamic scene with known ground-truth geometry and held-out cameras, compute UA-4DGS's per-pixel uncertainty map, and compare it with the per-pixel error on the held-out views; the premise fails if the map does not concentrate where error is concentrated, or if it is exactly zero in disoccluded regions that nonetheless show large error.

Watch

Extended reading notes

Core claim

The paper claims that in casually recorded monocular video, 4D Gaussian Splatting overfits training frames because standard regularizers are applied everywhere. Its central discovery is that a per-primitive uncertainty score, computed by summing each Gaussian's alpha-blending contribution across training pixels, can be rendered into a per-pixel 2D uncertainty map for any unseen view; gating diffusion-based and depth-smoothness losses by this map concentrates regularization on regions that training data constrain least. The paper further claims that SfM's failure to initialize primitives in fast-moving regions can be fixed by seeding extra Gaussians at dynamic pixels identified by scene flow, using estimated depth to place them in 3D. Together these components let 4DGS improve both novel-view quality and training-image reconstruction on DyCheck, and the uncertainty-gating idea transfers to few-shot static scenes.

Load-bearing premise

The method assumes that how often and how strongly a Gaussian primitive contributed to rendering training pixels measures how reliably its parameters are known, so that low-contribution regions are the ones needing extra priors.

Editorial extensions

If this is right

  • Casual handheld monocular video becomes a usable input for 4D Gaussian Splatting: on DyCheck, UA-4DGS raises masked PSNR from 14.14 to 15.25 over the 4DGS baseline, with consistent SSIM and LPIPS gains.
  • Uncertainty gating resolves the regularization trade-off: unweighted diffusion and TV losses improve test views but degrade training reconstruction, whereas the weighted versions improve both.
  • Dynamic region densification prevents wasteful primitive cloning and out-of-memory failures by initializing Gaussians where SfM provides no landmarks.
  • The uncertainty-aware regularization is not specific to dynamics: applying it to FSGS improves few-shot static reconstruction on LLFF with only three training views.
  • The method retains real-time rendering, running at 75 FPS on an RTX A5000, so the gains do not come at the cost of interactivity.

Reading between the lines

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

  • Editorial inference: the same contribution-based uncertainty gate could be attached to any auxiliary loss in sparse-view Gaussian Splatting, so the mechanism is not tied to diffusion or depth smoothness.
  • Editorial inference: because the uncertainty map is zero wherever no primitive is splatted, fully disoccluded pixels receive no regularization; adding a coverage-dependent floor or seeding those areas would test whether the map measures uncertainty or only visibility.
  • Editorial inference: since dynamic densification leans on external scene-flow and depth estimates, the method's gains in fast-moving regions should track estimator accuracy; corrupting those inputs would quantify that dependence.
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 proposes UA-4DGS, a 4D Gaussian Splatting method for dynamic scenes from casually recorded monocular videos. It introduces an uncertainty-aware regularization that estimates per-primitive uncertainty from each primitive's aggregate contribution to training-image rendering (Eqs. 7-8), renders this into a 2D uncertainty map for novel views (Eq. 9), and uses the map to weight a diffusion-based loss (Eq. 12) and a depth-smoothness total-variation loss (Eq. 14). It also proposes a dynamic-region densification step (Section 4.2) that initializes Gaussian primitives in fast-moving regions using estimated depth and scene flow. Experiments on the DyCheck dataset report improvements over D-3DGS, Zhan et al., and 4DGS in novel-view metrics, with a per-scene breakdown in the appendix. The method is additionally applied to few-shot static scene reconstruction on LLFF using FSGS as the base.

Significance. If the proposed mechanism is sound, the paper addresses a real and under-explored problem: 4D Gaussian Splatting overfits on casual monocular videos, and existing regularization techniques trade off training-image reconstruction quality against novel-view performance. The uncertainty-weighted regularization is a plausible way to reduce that trade-off, and the dynamic-region densification directly targets a failure mode (missing SfM initialization in fast-moving areas) that is not handled by prior 4DGS baselines. The paper also provides quantitative gains over strong baselines, with per-scene and ablative results. The main contributions are therefore potentially useful. However, the central uncertainty proxy is not externally validated, and the uncertainty map has a structural blind spot in unobserved regions, so the significance of the claimed gains is not yet established.

major comments (4)
  1. [Section 4.1, Eqs. (9), (12), (14)] The uncertainty map U(r) is rendered from the existing Gaussian primitives via the same alpha-blending weights used for the image. Consequently, at any pixel where no primitive projects, all omega_k(r)=0 and therefore U(r)=0; by Eqs. (12) and (14) both the diffusion loss and the depth-smoothness loss vanish exactly there. This is precisely the disoccluded, unobserved region that the paper motivates in Section 1, so the proposed regularization is structurally disabled where it is most needed. The paper's limitation paragraph only concedes image-level 2D maps; it does not address this zero-in-empty-region failure. Please either demonstrate empirically that such regions are nevertheless covered by other terms, or modify the regularization so that empty regions receive explicit uncertainty.
  2. [Section 4.1, Eqs. (7)-(8)] The proxy C_k, defined as the aggregate training-image contribution of a primitive, is asserted to measure parameter reliability, but no derivation or external validation is provided. A primitive can have high training contribution and still be poorly constrained at a novel pose or time because its deformation or view-dependent color is ambiguous, while a rarely used primitive can be well constrained by a few consistent observations. The paper needs a direct test of this proxy, e.g., correlating U(r) with the actual rendering error on held-out views, or an ablation that replaces the uncertainty weighting with a uniform weighting of the same losses. Without such evidence, the attributed gains could come from the data-driven losses and densification rather than from the uncertainty mechanism.
  3. [Section 5.3 and Appendix A.1, Tables 3 and 5] The ablation in Table 3 is conducted on a single DyCheck scene (spin), yet the method's main claim is a general improvement over seven scenes. The per-scene breakdown in Table 5 shows that the improvements are not consistent: on Apple the mSSIM gain over 4DGS is 0.001 (0.692 vs. 0.691), on Paper-windmill mSSIM is equal or slightly worse (0.207 vs. 0.208), and the teddy gain is small. The text in A.1 states 'consistent improvement,' which is not supported by these numbers. Please report ablations over all scenes, or justify why the single-scene ablation is representative.
  4. [Table 4 and Section 5.3] The claim that the method 'preserves reconstruction quality on training images' is only partially supported by Table 4. On the static room scene, training PSNR drops from 42.38 (FSGS) to 41.28 with L_UA-diff and to 38.26 with L_UA-TV; training LPIPS stays at 0.029, and SSIM improves slightly. The paper should either report training metrics on the dynamic DyCheck scenes or temper the claim, since the only direct training-reconstruction evidence shows a PSNR decrease.
minor comments (5)
  1. [Section 4, heading and Eq. (20)] The section heading 'Uncertanty-Aware 4D Gaussian Splatting' contains a typo; it should be 'Uncertainty-Aware.' In Eq. (20), the loss formula contains a double plus sign before lambda_UA-TV: '++lambda_UA-TV'.
  2. [Eq. (5)] In the transmittance product, the Gaussian kernel for the preceding primitive j should be evaluated at its own projected mean and covariance, i.e., G^pi_j(r; mu^pi_j, Sigma^pi_j), rather than at mu^pi_k and Sigma^pi_k as written.
  3. [Section 5.2, naming] The text says the method is called UA-FSGS when applied to FSGS, but Section 5.2 and Table 2 refer to it as UA-3DGS; please use one consistent name.
  4. [Appendix A.1, Table 5] In Table 5, the Paper-windmill row lists mPSNR 11.399 for D-3DGS, which has an extra decimal place and appears inconsistent with the other entries; please format all numbers to the same precision.
  5. [Appendix A.2] The phrase 'causally captured monocular videos' should be 'casually captured monocular videos.' Also, the statement that uncertainty-aware regularization starts at iteration 20,000 and that refined images are stored for 2,000 iterations would benefit from a brief explanation of how the 200 sampled images are selected.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the uncertainty map is a rendered heuristic from the model's own blending weights, not a prediction forced by its inputs.

full rationale

No circular step rises to the level of a derivation equated with its input. Eq. 7 defines C_k as the cumulative α-blending contribution to training pixels, Eq. 8 maps it to U_k, and Eq. 9 renders a 2D uncertainty map with the same α-blending weights. This is self-referential in the sense that the regularization weight is a function of the current model, but it is not equivalent to the target: L_UA-diff (Eq. 12) and L_UA-TV (Eq. 14) use U only to weight external diffusion and depth-smoothness priors, and no term reduces to the reconstruction loss or to the DyCheck metric. The hyperparameters (c0, c1, λ values) are hand-set in Appendix A.2; the phrase 'for optimal performance' for LLFF weights is ordinary hyperparameter selection, not a fitted variable relabeled as prediction. The only self-citation of consequence, InfoNeRF [25], appears in related work on ray-entropy uncertainty, while the paper explicitly uses contribution-based uncertainty instead, so no load-bearing self-citation chain exists. The paper's own limitations section concedes image-level 2D uncertainty maps, and the skeptical U=0-in-disoccluded-region issue follows structurally from Eq. 9; both are effectiveness/correctness limitations, not circularity. The method is evaluated against external DyCheck and LLFF baselines, and its central gains do not reduce to the paper's own definitions.

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

The central method rests on a hand-defined uncertainty proxy and on external depth, flow, and diffusion priors. There are no new physical entities. The main parameters are loss weights, sigmoid normalization constants, and the schedule for when regularization begins. The uncertainty proxy is the most consequential assumption because the whole regularization scheme inherits its failures.

free parameters (5)
  • Sigmoid inflection point c0 = 0.25
    Hand-chosen to shift the uncertainty normalization for C_k in Eq. 8; affects which primitives are treated as uncertain.
  • Sigmoid slope c1 = 20/L
    Hand-chosen slope depending on number of training images L; controls the sharpness of the uncertainty threshold.
  • Loss weights lambda_data, lambda_UA-diff, lambda_UA-TV (DyCheck) = 0.5, 0.2, 0.01
    Balancing hyperparameters in Eq. 20; chosen by hand for the DyCheck experiments.
  • Loss weights lambda_UA-diff, lambda_UA-TV (LLFF) = 0.1, 0.001
    Chosen separately for the few-shot static experiments, described in Appendix A.2 as selected for optimal performance.
  • Uncertainty regularization start iteration = 20,000
    Uncertainty-aware losses are applied only after iteration 20,000 of 40,000; this is a hand-set schedule choice.
assumptions (4)
  • domain assumption Aggregate training contribution C_k, as a sum of alpha-blending weights, is a valid proxy for parameter estimation confidence.
    Section 4.1, Eq. 7 to 8. The paper asserts that informative primitives are accurately estimated and uninformative ones are uncertain, without a statistical derivation.
  • domain assumption Alpha-blending primitive uncertainties produces a meaningful uncertainty map for unseen views, including coverage of genuinely unobserved regions.
    Section 4.1, Eq. 9. If no primitive projects to a pixel, U(r) is zero, so disoccluded empty regions receive no regularization.
  • domain assumption DDIM-refined images IDDIM are sufficiently aligned with rendered images for pixel-wise loss computation.
    Section 4.1 after Eq. 11. The paper acknowledges IDDIM sometimes contains misaligned context, yet the uncertainty-weighted loss still assumes spatial alignment.
  • domain assumption Scene flow and depth estimates from [71] are reliable enough to identify dynamic pixels and initialize Gaussian primitives at correct 3D locations.
    Section 4.2, Eq. 15. Incorrect depth or flow would place new primitives in wrong positions, and the method has no temporal consistency check.

how reviews work

0 comments
Cite this review

Pith. "Pith review of 4D Gaussian Splatting in the Wild with Uncertainty-Aware Regularization." pith.science (2026). https://pith.science/paper/QHIVHONG

@misc{pith2026241108879,
  author       = {Pith},
  title        = {Pith review of: 4D Gaussian Splatting in the Wild with Uncertainty-Aware Regularization},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/QHIVHONG}},
  note         = {Machine review of arXiv:2411.08879}
}
read the original abstract

Novel view synthesis of dynamic scenes is becoming important in various applications, including augmented and virtual reality. We propose a novel 4D Gaussian Splatting (4DGS) algorithm for dynamic scenes from casually recorded monocular videos. To overcome the overfitting problem of existing work for these real-world videos, we introduce an uncertainty-aware regularization that identifies uncertain regions with few observations and selectively imposes additional priors based on diffusion models and depth smoothness on such regions. This approach improves both the performance of novel view synthesis and the quality of training image reconstruction. We also identify the initialization problem of 4DGS in fast-moving dynamic regions, where the Structure from Motion (SfM) algorithm fails to provide reliable 3D landmarks. To initialize Gaussian primitives in such regions, we present a dynamic region densification method using the estimated depth maps and scene flow. Our experiments show that the proposed method improves the performance of 4DGS reconstruction from a video captured by a handheld monocular camera and also exhibits promising results in few-shot static scene reconstruction.

Figures

Figures reproduced from arXiv: 2411.08879 by the authors.

Figure 1
Figure 1. Concept of uncertainty-aware regularization. Existing models often use regularization [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Visualization of the dynamic region densification on the [PITH_FULL_IMAGE:figures/full_fig_p007_2.png] view at source ↗
Figure 3
Figure 3. Qualitative results on the space-out, paper-windmill, teddy, and spin scenes in the DyCheck dataset. UA-4DGS (Ours) shows outstanding quality of rendered images compared to existing methods, including D-3DGS [68], Zhan et al. [31], and 4DGS [61] 10 [PITH_FULL_IMAGE:figures/full_fig_p010_3.png] view at source ↗
Figures from the paper (2 more)
Figure 4
Figure 4. Figure 4: Qualitative comparison between UA-4DGS and other methods tested on the DyCheck [PITH_FULL_IMAGE:figures/full_fig_p016_4.png]
Figure 5
Figure 5. Figure 5: Qualitative comparison between UA-4DGS and other methods tested on the DyCheck [PITH_FULL_IMAGE:figures/full_fig_p017_5.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

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

  1. Generative Physical AI in Vision: A Survey

    cs.CV 2025-01 conditional novelty 6.0 of 10

    A structured review that categorizes physics-aware generative models in vision into explicit-simulation and implicit-learning families and proposes six integration paradigms.

Reference graph

Works this paper leans on

72 extracted references · 59 canonical work pages · cited by 1 Pith paper

  1. [1]

    J. An. Model development with vessl, 2023. URL https://www.vessl.ai/. Software available from vessl.ai

    work page 2023

  2. [2]

    J. Bai, L. Huang, W. Gong, J. Guo, and Y . Guo. Self-nerf: A self-training pipeline for few-shot neural radiance fields. arXiv:2303.05775, 2023

    work page Pith review arXiv 2023

  3. [3]

    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. In ICCV, 2021

    2021

  4. [4]

    Bonotto, L

    M. Bonotto, L. Sarrocco, D. Evangelista, M. Imperoli, and A. Pretto. Combinerf: A combination of regularization techniques for few-shot neural radiance field view synthesis.arXiv:2403.14412, 2024

    arXiv 2024

  5. [5]

    Cao and J

    A. Cao and J. Johnson. Hexplane: A fast representation for dynamic scenes. In CVPR, 2023

    work page 2023

  6. [6]

    A. Chen, Z. Xu, A. Geiger, J. Yu, and H. Su. Tensorf: Tensorial radiance fields. In ECCV, 2022

    2022

  7. [7]

    Chung, J

    J. Chung, J. Oh, and K. M. Lee. Depth-regularized optimization for 3d gaussian splatting in few-shot images. arXiv:2311.13398, 2023

    arXiv 2023

  8. [8]

    K. Deng, A. Liu, J.-Y . Zhu, and D. Ramanan. Depth-supervised nerf: Fewer views and faster training for free. In CVPR, 2022

    work page 2022

Show all 72 references
  1. [9]

    Doersch, Y

    C. Doersch, Y . Yang, M. Vecerik, D. Gokay, A. Gupta, Y . Aytar, J. Carreira, and A. Zisserman. Tapir: Tracking any point with per-frame initialization and temporal refinement. In ICCV, 2023

    2023

  2. [10]

    Y . Du, Y . Zhang, H.-X. Yu, J. B. Tenenbaum, and J. Wu. Neural radiance flow for 4d view synthesis and video processing. In ICCV, 2021

    2021

  3. [11]

    J. Fang, T. Yi, X. Wang, L. Xie, X. Zhang, W. Liu, M. Nießner, and Q. Tian. Fast dynamic radiance fields with time-aware neural voxels. In SIGGRAPH Asia, 2022

    2022

  4. [12]

    Fridovich-Keil, G

    S. Fridovich-Keil, G. Meanti, F. R. Warburg, B. Recht, and A. Kanazawa. K-planes: Explicit radiance fields in space, time, and appearance. In CVPR, 2023

    2023

  5. [13]

    C. Gao, A. Saraf, J. Kopf, and J.-B. Huang. Dynamic view synthesis from dynamic monocular video. In ICCV, 2021

    2021

  6. [14]

    H. Gao, R. Li, S. Tulsiani, B. Russell, and A. Kanazawa. Monocular dynamic view synthesis: A reality check. In NeurIPS, 2022

    2022

  7. [15]

    S. J. Garbin, M. Kowalski, M. Johnson, J. Shotton, and J. Valentin. Fastnerf: High-fidelity neural rendering at 200fps. In ICCV, 2021

    2021

  8. [16]

    S. Guo, Q. Wang, Y . Gao, R. Xie, and L. Song. Depth-guided robust and fast point cloud fusion nerf for sparse input views. In AAAI, 2024

    2024

  9. [17]

    S. Hu, K. Zhou, K. Li, L. Yu, L. Hong, T. Hu, Z. Li, G. H. Lee, and Z. Liu. Consistentnerf: En- hancing neural radiance fields with 3d consistency for sparse view synthesis. arXiv:2305.11031, 2023

    2023 arXiv

  10. [18]

    Huang, Y .-T

    Y .-H. Huang, Y .-T. Sun, Z. Yang, X. Lyu, Y .-P. Cao, and X. Qi. Sc-gs: Sparse-controlled gaussian splatting for editable dynamic scenes. In CVPR, 2024. 11

    2024

  11. [19]

    Hyung, S

    J. Hyung, S. Hong, S. Hwang, J. Lee, J. Choo, and J.-H. Kim. Effective rank analysis and regularization for enhanced 3d gaussian splatting. arXiv preprint arXiv:2406.11672, 2024

    2024 arXiv

  12. [20]

    A. Jain, M. Tancik, and P. Abbeel. Putting nerf on a diet: Semantically consistent few-shot view synthesis. In ICCV, 2021

    2021

  13. [21]

    Jiang, B

    W. Jiang, B. Lei, and K. Daniilidis. Fisherrf: Active view selection and uncertainty quantification for radiance fields using fisher information. In ECCV, 2024

    2024

  14. [22]

    Kanaoka, M

    D. Kanaoka, M. Sonogashira, H. Tamukoh, and Y . Kawanishi. Manifoldnerf: View-dependent image feature supervision for few-shot neural radiance fields. arXiv:2310.13670, 2023

    2023 arXiv

  15. [23]

    Kerbl, G

    B. Kerbl, G. Kopanas, T. Leimkühler, and G. Drettakis. 3d gaussian splatting for real-time radiance field rendering. In ACM ToG, 2023

    2023

  16. [24]

    I. Kim, M. Choi, and H. J. Kim. Up-nerf: Unconstrained pose prior-free neural radiance field. In NeurIPS, 2023

    2023

  17. [25]

    M. Kim, S. Seo, and B. Han. Infonerf: Ray entropy minimization for few-shot neural volume rendering. In CVPR, 2022

    2022

  18. [26]

    M.-S. Kwak, J. Song, and S. Kim. Geconerf: Few-shot neural radiance fields via geometric consistency. In ICML, 2023

    2023

  19. [27]

    S. Lee, L. Chen, J. Wang, A. Liniger, S. Kumar, and F. Yu. Uncertainty guided policy for active robotic 3d reconstruction using neural radiance fields. IEEE Robotics and Automation Letters , 2022

    2022

  20. [28]

    J. Li, D. Li, C. Xiong, and S. Hoi. Blip: Bootstrapping language-image pre-training for unified vision-language understanding and generation. In ICML, 2022

    2022

  21. [29]

    T. Li, M. Slavcheva, M. Zollhoefer, S. Green, C. Lassner, C. Kim, T. Schmidt, S. Lovegrove, M. Goesele, R. Newcombe, et al. Neural 3d video synthesis from multi-view video. In CVPR, 2022

    2022

  22. [30]

    Z. Li, S. Niklaus, N. Snavely, and O. Wang. Neural scene flow fields for space-time view synthesis of dynamic scenes. In CVPR, 2021

    2021

  23. [31]

    Z. Li, Z. Chen, Z. Li, and Y . Xu. Spacetime gaussian feature splatting for real-time dynamic view synthesis. In CVPR, 2024

    2024

  24. [32]

    C. Meng, Y . He, Y . Song, J. Song, J. Wu, J.-Y . Zhu, and S. Ermon. Sdedit: Guided image synthesis and editing with stochastic differential equations. In ICLR, 2022

    2022

  25. [33]

    Mildenhall, P

    B. Mildenhall, P. P. Srinivasan, R. Ortiz-Cayon, N. K. Kalantari, R. Ramamoorthi, R. Ng, and A. Kar. Local light field fusion: Practical view synthesis with prescriptive sampling guidelines. In SIGGRAPH, 2019

    2019

  26. [34]

    Mildenhall, P

    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. In ECCV, 2020

    2020

  27. [35]

    Müller, A

    T. Müller, A. Evans, C. Schied, and A. Keller. Instant neural graphics primitives with a multiresolution hash encoding. ACM TOG, 2022

    2022

  28. [36]

    Niemeyer, J

    M. Niemeyer, J. T. Barron, B. Mildenhall, M. S. Sajjadi, A. Geiger, and N. Radwan. Regnerf: Regularizing neural radiance fields for view synthesis from sparse inputs. In CVPR, 2022

    2022

  29. [37]

    X. Pan, Z. Lai, S. Song, and G. Huang. Activenerf: Learning where to see with uncertainty estimation. In ECCV, 2022

    2022

  30. [38]

    K. Park, U. Sinha, J. T. Barron, S. Bouaziz, D. B. Goldman, S. M. Seitz, and R. Martin-Brualla. Nerfies: Deformable neural radiance fields. ICCV, 2021

    2021

  31. [39]

    K. Park, U. Sinha, P. Hedman, J. T. Barron, S. Bouaziz, D. B. Goldman, R. Martin-Brualla, and S. M. Seitz. Hypernerf: A higher-dimensional representation for topologically varying neural radiance fields. ACM Trans. Graph., 2021. 12

    2021

  32. [40]

    Paszke, S

    A. Paszke, S. Gross, F. Massa, A. Lerer, J. Bradbury, G. Chanan, T. Killeen, Z. Lin, N. Gimelshein, L. Antiga, et al. Pytorch: An imperative style, high-performance deep learning library, 2019

    2019

  33. [41]

    Pumarola, E

    A. Pumarola, E. Corona, G. Pons-Moll, and F. Moreno-Noguer. D-nerf: Neural radiance fields for dynamic scenes. In CVPR, 2021

    2021

  34. [42]

    G. Qian, J. Mai, A. Hamdi, J. Ren, A. Siarohin, B. Li, H.-Y . Lee, I. Skorokhodov, P. Wonka, S. Tulyakov, et al. Magic123: One image to high-quality 3d object generation using both 2d and 3d diffusion priors. In ICLR, 2024

    2024

  35. [43]

    Roessle, J

    B. Roessle, J. T. Barron, B. Mildenhall, P. P. Srinivasan, and M. Nießner. Dense depth priors for neural radiance fields from sparse input views. In CVPR, 2022

    2022

  36. [44]

    Rombach, A

    R. Rombach, A. Blattmann, D. Lorenz, P. Esser, and B. Ommer. High-resolution image synthesis with latent diffusion models. In CVPR, 2022

    2022

  37. [45]

    N. Ruiz, Y . Li, V . Jampani, Y . Pritch, M. Rubinstein, and K. Aberman. Dreambooth: Fine tuning text-to-image diffusion models for subject-driven generation. In CVPR, 2023

    2023

  38. [46]

    Savant, D

    L. Savant, D. Valsesia, and E. Magli. Modeling uncertainty for gaussian splatting. arXiv preprint arXiv:2403.18476, 2024

    2024 arXiv

  39. [47]

    J. L. Schonberger and J.-M. Frahm. Structure-from-motion revisited. In CVPR, 2016

    2016

  40. [48]

    S. Seo, Y . Chang, and N. Kwak. Flipnerf: Flipped reflection rays for few-shot novel view synthesis. In ICCV, 2023

    2023

  41. [49]

    S. Seo, D. Han, Y . Chang, and N. Kwak. Mixnerf: Modeling a ray with mixture density for novel view synthesis from sparse inputs. In CVPR, 2023

    2023

  42. [50]

    R. Shao, Z. Zheng, H. Tu, B. Liu, H. Zhang, and Y . Liu. Tensor4d: Efficient neural 4d decomposition for high-fidelity dynamic reconstruction and rendering. In CVPR, 2023

    2023

  43. [51]

    J. Shen, A. Ruiz, A. Agudo, and F. Moreno-Noguer. Stochastic neural radiance fields: Quantify- ing uncertainty in implicit 3d representations. In 3DV, 2021

    2021

  44. [52]

    J. Shen, A. Agudo, F. Moreno-Noguer, and A. Ruiz. Conditional-flow nerf: Accurate 3d modelling with reliable uncertainty quantification. In ECCV, 2022

    2022

  45. [53]

    Somraj and R

    N. Somraj and R. Soundararajan. Vip-nerf: Visibility prior for sparse input neural radiance fields. In SIGGRAPH Conference, 2023

    2023

  46. [54]

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

    2010 arXiv

  47. [55]

    J. Song, S. Park, H. An, S. Cho, M.-S. Kwak, S. Cho, and S. Kim. Därf: Boosting radiance fields from sparse input views with monocular depth adaptation. In NeurIPs, 2023

    2023

  48. [56]

    Sünderhauf, J

    N. Sünderhauf, J. Abou-Chakra, and D. Miller. Density-aware nerf ensembles: Quantifying predictive uncertainty in neural radiance fields. In ICRA, 2023

    2023

  49. [57]

    Teed and J

    Z. Teed and J. Deng. Raft: Recurrent all-pairs field transforms for optical flow. In ECCV, 2020

    2020

  50. [58]

    Truong, M.-J

    P. Truong, M.-J. Rakotosaona, F. Manhardt, and F. Tombari. Sparf: Neural radiance fields from sparse and noisy poses. In CVPR, 2023

    2023

  51. [59]

    G. Wang, Z. Chen, C. C. Loy, and Z. Liu. Sparsenerf: Distilling depth ranking for few-shot novel view synthesis. In CVPR, 2023

    2023

  52. [60]

    L. Wang, J. Zhang, X. Liu, F. Zhao, Y . Zhang, Y . Zhang, M. Wu, J. Yu, and L. Xu. Fourier plenoctrees for dynamic radiance field rendering in real-time. In CVPR, 2022

    2022

  53. [61]

    G. Wu, T. Yi, J. Fang, L. Xie, X. Zhang, W. Wei, W. Liu, Q. Tian, and X. Wang. 4d gaussian splatting for real-time dynamic scene rendering. In CVPR, 2024. 13

    2024

  54. [62]

    Wynn and D

    J. Wynn and D. Turmukhambetov. Diffusionerf: Regularizing neural radiance fields with denoising diffusion models. In CVPR, 2023

    2023

  55. [63]

    Y . Xiao, X. Liu, D. Zhai, K. Jiang, J. Jiang, and X. Ji. Sgcnerf: Few-shot neural rendering via sparse geometric consistency guidance. arXiv:2404.00992, 2024

    2024 arXiv

  56. [64]

    T. Xie, Z. Zong, Y . Qiu, X. Li, Y . Feng, Y . Yang, and C. Jiang. Physgaussian: Physics-integrated 3d gaussians for generative dynamics. In CVPR, 2024

    2024

  57. [65]

    D. Xu, Y . Jiang, P. Wang, Z. Fan, H. Shi, and Z. Wang. Sinnerf: Training neural radiance fields on complex scenes from a single image. In ECCV, 2022

    2022

  58. [66]

    D. Yan, J. Liu, F. Quan, H. Chen, and M. Fu. Active implicit object reconstruction using uncertainty-guided next-best-view optimization. IEEE Robotics and Automation Letters , 2023

    2023

  59. [67]

    J. Yang, M. Pavone, and Y . Wang. Freenerf: Improving few-shot neural rendering with free frequency regularization. In CVPR, 2023

    2023

  60. [68]

    Z. Yang, X. Gao, W. Zhou, S. Jiao, Y . Zhang, and X. Jin. Deformable 3d gaussians for high-fidelity monocular dynamic scene reconstruction. In CVPR, 2024

    2024

  61. [69]

    J. S. Yoon, K. Kim, O. Gallo, H. S. Park, and J. Kautz. Novel view synthesis of dynamic scenes with globally coherent depths from a monocular camera. In CVPR, 2020

    2020

  62. [70]

    H. Zhan, J. Zheng, Y . Xu, I. Reid, and H. Rezatofighi. Activermap: Radiance field for active mapping and planning. arXiv:2211.12656, 2022

    2022 arXiv

  63. [71]

    Zhang, F

    Z. Zhang, F. Cole, R. Tucker, W. T. Freeman, and T. Dekel. Consistent depth of moving objects in video. ACM TOG, 2021

    2021

  64. [72]

    Z. Zhu, Z. Fan, Y . Jiang, and Z. Wang. Fsgs: Real-time few-shot view synthesis using gaussian splatting. 2024. 14 A Appendix A.1 Per-Scene Breakdown DyCheck dataset Supplementing Table 1 of the main paper, we show the experimental results from individual scenes in terms of mP...

    2024

Pith tools

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