Pith. sign in

REVIEW 4 major objections 5 minor 51 references

RegGS: Unposed Sparse Views Gaussian Splatting with 3DGS Registration

T0 review · 4 major / 5 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read RegGS reconstructs scenes from sparse unposed images by registering feed-forward 3D Gaussian maps with optimal transport, reporting precise poses and high-quality novel views.

desk verdict A plausible engineering integration for registering feed-forward 3D Gaussians beyond two views, but the main experimental tables contradict the ablation table on the same 16-view RE10K configuration, so the reported gains are not yet internally consistent. read the letter →

arxiv 2507.08136 v2 pith:AHGC36FX submitted 2025-07-10 cs.CV

Chong Cheng , Yu Hu , Sicheng Yu , Beizhen Zhao , Zijian Wang , Hao Wang This is my paper
classification cs.CV
keywords 3DGaussianSplattingunposedsparseviewreconstructionmixturemodelregistration2-WassersteindistanceSinkhornoptimaltransportSim(3)alignmentcameraposeestimationnovelsynthesis
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

RegGS tackles the problem of reconstructing a 3D scene from a handful of images with no known camera poses. The paper's claim is that instead of optimising a global scene directly, you can let a pretrained feed-forward network predict local 3D Gaussian maps for each input and then register those maps into one consistent representation. The registration is driven by a Mixture 2-Wasserstein distance between Gaussian mixture models, computed efficiently with an entropy-regularized Sinkhorn algorithm, plus photometric and depth losses. Concretely, on RE10K with 16 views it reports PSNR 28.663 and an ATE of 0.041, better than the unposed baselines it compares against.

What carries the argument

The load-bearing object is the Mixture 2-Wasserstein ($\mathrm{MW}_2$) distance between two Gaussian mixture models, computed with an entropy-regularized Sinkhorn algorithm. This differentiable metric measures how far the transformed sub-map's Gaussians are from the main map's Gaussians, including both means and covariances, and it is what lets the registration solve for a $\mathrm{Sim}(3)$ transformation. The joint registration module then combines this structural term with rendered photometric and depth losses, refining the coarse optimal-transport alignment into precise geometry.

What would settle it

Take a scene dominated by a reflective or textureless surface, the failure mode shown in Fig. 9: if the upstream model misplaces the depth of those Gaussians, the MW2, photometric, and depth losses should fail to register the sub-map, and ATE and PSNR should degrade sharply even though the same method succeeds on diffuse scenes. A quantitative version is to compare this ATE gap to the upstream depth error on the same surfaces.

Watch

Extended reading notes

Core claim

The paper claims that local 3D Gaussian maps, each generated from a small number of images by a pretrained feed-forward model, can be fused into a single globally consistent 3D Gaussian scene by solving for the similarity transformation (rotation, translation, and uniform scale) in $\mathrm{Sim}(3)$ that minimizes an optimal-transport distance between the maps treated as Gaussian mixture models. It further claims that combining this structural distance with photometric and depth consistency losses in a differentiable joint registration module yields accurate camera poses and high-quality novel views from unposed sparse images, demonstrated on RE10K and ACID. The method is incremental: two images seed a main map, each new view contributes a sub-map that is aligned and merged, and a final global refinement polishes the result.

Load-bearing premise

The pipeline rests on the pretrained feed-forward model producing local 3D Gaussians that are structurally reliable enough that a single similarity transform can align each sub-map with the main map; if any sub-map's geometry is poor, the registration losses cannot repair it.

Editorial extensions

If this is right

  • The same registered map directly yields a camera pose for every input view, removing the need for a separate structure-from-motion stage.
  • The approach is not capped at the two-view limit of feed-forward models: as more views arrive, each new local map is merged into the growing global map and a final refinement polishes the whole scene.
  • Because the structural MW2 term ignores colour, the alignment stage is insensitive to lighting changes across views, while the photometric and depth terms then refine local geometry.
  • Removing any one of the three losses sharply degrades both pose and rendering quality, so the joint registration design is the load-bearing contribution.

Reading between the lines

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

  • The same optimal-transport registration could be applied to align sub-maps from other feed-forward 3D predictors or even from NeRF-style reconstructions whose output has been converted to Gaussian components.
  • A practical next step is to warm-start the Sinkhorn transport plan from the previous frame during incremental scanning, which would lower the reported runtime bottleneck without changing the method.
  • The documented failure on reflective surfaces suggests that training the upstream generator with explicit depth supervision on specular geometry would likely reduce registration failures more than improving the registration module itself.
  • The large pose-accuracy gains at 8 views point toward sequential SLAM and embodied-agent settings, where poses must be recovered online from a stream of sparse views.
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 RegGS, a framework for reconstructing 3D Gaussian scenes from unposed sparse views by incrementally registering local 3D Gaussian maps produced by a pre-trained feed-forward model (NoPoSplat). The alignment is performed in Sim(3) space using an entropy-regularized Sinkhorn approximation of the Mixture 2-Wasserstein (MW2) distance between Gaussian mixture models, combined with photometric and depth-consistency losses in a joint differentiable registration module, followed by global refinement. The paper reports state-of-the-art novel-view synthesis and pose-estimation results on RE10K and ACID, with an ablation study of the loss components and the registration module. The appendix provides a claimed gradient-consistency proof for the entropy-regularized MW2 distance and a complexity analysis of the Sinkhorn algorithm.

Significance. If the reported results are correct, RegGS is a practically valuable contribution: it extends feed-forward Gaussian predictors beyond their fixed input-size limitation, uses a differentiable optimal-transport metric for Gaussian-mixture alignment, and shows large improvements over unposed optimization-based baselines on two benchmarks. The paper includes useful qualitative comparisons, runtime/memory measurements, and an honest discussion of failure modes. However, the central empirical evidence contains an unresolved internal inconsistency between the main results and the ablation tables, and the theoretical justification for the gradient-consistency claim is not rigorous. These issues need to be addressed before the main claims can be fully accepted.

major comments (4)
  1. [Sec. 4.3, Tables 1, 3, and 4] There is an internal inconsistency in the reported 16-view RE10K results. Table 3 reports RegGS with ATE 0.041 at 16 views, and Table 1 reports PSNR 28.663 for the same setting. However, Table 4, which states 'the input is set to 16 views' on RE10K and evaluates the full model ('RegGS (Ours)'), reports ATE 0.098 and PSNR 23.09. Unless the ablation is run under a different protocol (no global refinement, different split, different seed, or a subset of scenes), these numbers are irreconcilable. Please state the exact protocol for the ablation experiments and explain the discrepancy. Because the central claims of precise pose estimation and high-quality novel-view synthesis rest on Tables 1 and 3, this inconsistency undermines the empirical support and must be resolved.
  2. [Supplementary Sec. 6, Eqs. (19)-(25)] The gradient-consistency proof is not a valid proof. Gamma-convergence of the functionals in Eq. (20) concerns convergence of minimizers and does not imply convergence of gradients; Eq. (22) invokes an implicit-function/envelope-theorem argument without verifying the required differentiability and regularity conditions for the entropic optimal transport plan pi*_ik(epsilon); and Eq. (23) asserts convergence of the regularized plans to an optimal plan, which is not established by Gamma-convergence. Consequently, the claimed limit in Eq. (25), lim_{epsilon->0} grad W2,epsilon in subdifferential W2, is unsupported. Since this claim is used to justify the decreasing-epsilon optimization schedule, the proof should either be made rigorous or the theoretical statement should be substantially softened.
  3. [Sec. 4.1 and Tables 4-5] The MW2 distance is used both as the alignment objective in Eq. (13) and as the '3DGS registration evaluation' metric in Sec. 4.1 and in the ablation tables. This creates a circularity: the MW2 column in the ablations largely measures how well the trained loss is minimized, not an independent geometric quality. The ATE and NVS metrics are independent and mitigate this, but the registration-quality claim should be validated with an independent geometric measure (for example, Chamfer distance between point clouds or rendered depth maps), or the MW2 column should be presented explicitly as a loss-curve indicator rather than as an evaluation metric.
  4. [Sec. 4.1, Implementation Details] The paper does not report several free parameters that are essential for reproducibility: the loss weights lambda1, lambda2, lambda3 in Eq. (13), the entropy regularization parameter epsilon, the number of Sinkhorn iterations T, the optimizer settings (learning rate, number of iterations), and the specifics of scale normalization. An experimental paper whose central claim is a new optimization-based pipeline should provide these values in the main text or supplementary material. Please add a complete implementation-details subsection.
minor comments (5)
  1. [Supplementary Sec. 7] The first paragraph says the main map and submap contain 'M and N Gaussian gradients'; this should be 'Gaussian components' or 'Gaussians'.
  2. [Sec. 3.2] The phrase 'the two sets of Gaussian spheres are not in one-to-one correspondence' should be rephrased to 'Gaussian components' for consistency with the rest of the paper.
  3. [Sec. 4.1, Baselines] The baseline name is written as 'MARSt3R [24]' in the text but as 'MASt3R' in Tables 1 and 2; please make the notation consistent.
  4. [Supplementary Sec. 8] The text says 'As shown in Tab. 7, RegGS maintains controlled memory usage', but the corresponding item is presented as Figure 7; the cross-reference should be corrected.
  5. [Sec. 4.4 and Supplementary Fig. 9] The limitation discussion would be clearer if it explicitly referenced the supplementary failure example, since the main text's reference to reflective-surface failures is otherwise not tied to a figure.

Circularity Check

1 steps flagged · score 4.0 of 10

Registration-quality evaluation reuses the optimized MW2 objective as the metric; central NVS and pose claims are independently grounded.

  1. self definitional [Sec. 3.3 Eq. (13) and Sec. 4.1 'Evaluation Metrics']
    "Ltotal = λ1LMW2 + λ2LPhoto + λ3LDepth, (13) where the global alignment term LMW2 = W^2_{2,ϵ}(GA, T(GB)) is calculated using the differentiable Sinkhorn algorithm from Sec. 3.2... For 3DGS registration evaluation, we use the MW2 distance."

    The quantity minimized as the MW2 loss in Eq. (13) is the same MW2 distance that Table 4 and Sec. 4.1 report as evidence of successful 3DGS registration. The ablation 'w/o MW2' shows MW2 rising to 67.7 from 56.5, i.e., the evaluation directly measures the optimized objective rather than an independent structural metric. Thus the claim that the joint registration improves alignment is, for this column, true by construction. The paper's central NVS and pose claims (PSNR/SSIM/LPIPS and ATE) are measured with external ground truth and remain independent.

full rationale

The main derivation chain—feed-forward Gaussians from NoPoSplat, entropy-regularized Sinkhorn MW2, joint Sim(3) registration, and global refinement—does not reduce to fitted parameters or to self-citation: the transformation T is optimized against the scene via Eq. (13), and pose ATE and NVS metrics are compared against external ground truth. References [10, 11, 44] are prior-work context and are not load-bearing for the central result. The only concrete circularity is that the '3DGS registration evaluation' metric is the MW2 distance, which is exactly the LMW2 term being minimized, making Table 4's MW2 comparisons and the Fig. 5 claim that MW2 'precisely quantifies proximity' self-referential rather than independent validation. Separately, the reported 16-view RE10K numbers differ between Tables 1/3 (PSNR 28.663, ATE 0.041) and Table 4 (PSNR 23.09, ATE 0.098) for the same 'RegGS (Ours)' configuration; this is an internal-consistency and correctness concern, not a circularity. Overall circularity is moderate and confined to the registration-quality metric; the principal empirical claims retain independent support.

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

No new physical entities are introduced. The main assumptions are inherited from optimal transport theory and the upstream 3DGS model, plus standard domain assumptions about known intrinsics and rigid-ish alignment.

free parameters (5)
  • λ1, λ2, λ3 loss weights = not reported
    Weights in Eq. 13 balance MW2, photometric, and depth losses; their values are not given and likely tuned on validation/test data.
  • ε (entropy regularization) = not reported
    Controls Sinkhorn smoothness in Eq. 7; value not disclosed.
  • T (Sinkhorn iterations) = not reported
    Number of scaling updates in Eq. 8; text says T≤50 typically, but exact value not given.
  • Joint optimization hyperparameters (learning rate, number of iterations) = not reported
    Not specified; required to reproduce the reported ATE and NVS numbers.
  • Scale normalization average depth = not reported
    Sec. 3.4 uses average depth of sub Gaussians to normalize scale; computation detail omitted.
assumptions (5)
  • standard math Standard optimal transport theory: MW2 distance between Gaussian mixtures and Sinkhorn approximation.
    Used in Sec. 3.2 to define the alignment metric and algorithm.
  • ad hoc to paper Γ-convergence and gradient consistency of entropy-regularized Wasserstein distances (Eq. 19-25 in supplementary).
    The paper asserts that the regularized gradient converges to a subgradient of the exact W2; the proof is sketchy and not fully established, but the claim is used to justify optimization.
  • domain assumption The feed-forward network (NoPoSplat) produces Gaussian components whose weights sum to 1 after opacity normalization and whose covariances are valid positive-definite matrices.
    Sec. 3.1 Eq. 1-2 treat the output as a GMM; the method's correctness depends on this.
  • domain assumption Known camera intrinsics.
    The datasets provide intrinsics; the feed-forward model and rendering pipeline require them, but the paper does not discuss estimating intrinsics.
  • domain assumption A single Sim(3) transform is sufficient to align each sub-map to the main map.
    Sec. 3.3 models alignment in Sim(3); non-rigid distortions between local maps are not handled.

how reviews work

0 comments
Cite this review

Pith. "Pith review of RegGS: Unposed Sparse Views Gaussian Splatting with 3DGS Registration." pith.science (2026). https://pith.science/paper/AHGC36FX

@misc{pith2026250708136,
  author       = {Pith},
  title        = {Pith review of: RegGS: Unposed Sparse Views Gaussian Splatting with 3DGS Registration},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/AHGC36FX}},
  note         = {Machine review of arXiv:2507.08136}
}
abstract

3D Gaussian Splatting (3DGS) has demonstrated its potential in reconstructing scenes from unposed images. However, optimization-based 3DGS methods struggle with sparse views due to limited prior knowledge. Meanwhile, feed-forward Gaussian approaches are constrained by input formats, making it challenging to incorporate more input views. To address these challenges, we propose RegGS, a 3D Gaussian registration-based framework for reconstructing unposed sparse views. RegGS aligns local 3D Gaussians generated by a feed-forward network into a globally consistent 3D Gaussian representation. Technically, we implement an entropy-regularized Sinkhorn algorithm to efficiently solve the optimal transport Mixture 2-Wasserstein $(\text{MW}_2)$ distance, which serves as an alignment metric for Gaussian mixture models (GMMs) in $\mathrm{Sim}(3)$ space. Furthermore, we design a joint 3DGS registration module that integrates the $\text{MW}_2$ distance, photometric consistency, and depth geometry. This enables a coarse-to-fine registration process while accurately estimating camera poses and aligning the scene. Experiments on the RE10K and ACID datasets demonstrate that RegGS effectively registers local Gaussians with high fidelity, achieving precise pose estimation and high-quality novel-view synthesis. Project page: https://3dagentworld.github.io/reggs/.

Figures

Figures reproduced from arXiv: 2507.08136 by the authors.

Figure 1
Figure 1. Overview of our pipeline for 3D Gaussian Splatting from multiple unposed sparse views. A pre-trained feed-forward GS model [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Pipeline of Unposed Sparse Views Gaussian Splatting with 3DGS Registration (RegGS). First, we use a pre-trained feed￾forward Gaussian model to construct a main Gaussians from two initial images. Then, for each new input, a sub Gaussians is generated and aligned with the main Gaussians. Specifically, by solving the optimal transport MW2 distance with an entropy-regularized Sinkhorn approximation, our differentiable 3… view at source ↗
Figure 3
Figure 3. Qualitative Comparison on the RE10K [46]. NoPoSplat: 2× views; others: 16× views. Our method not only registers the 3D Gaussians but also enhances novel view synthesis through global refinement. Method 2× 8× 16× 32× PSNR↑ SSIM↑ LPIPS↓ PSNR↑ SSIM↑ LPIPS↓ PSNR↑ SSIM↑ LPIPS↓ PSNR↑ SSIM↑ LPIPS↓ COLMAP* [33] 9.687 0.266 0.533 7.171 0.135 0.676 18.904 0.614 0.294 22.911 0.725 0.219 Splatt3R [35] 13.951 0.442 0.443 - - - -… view at source ↗
Figures from the paper (5 more)
Figure 4
Figure 4. Figure 4: Qualitative comparison on the ACID [26]. NoPoSplat: 2x views; others: 16x views. Our method is applicable to both indoor scenes and drone-captured videos, demonstrating superior novel view synthesis performance. Method 2× 8× 16× 32× PSNR↑ SSIM↑ LPIPS↓ PSNR↑ SSIM↑ LPIPS…
Figure 5
Figure 5. Figure 5: MW2 distances effectively quantify alignment levels be￾tween sets of 3D Gaussians under various conditions. Notably, the rightmost case aligns with the correct position. [14], CF-3DGS [20], MARSt3R [24], and Splatt3R [35]. Implementation Details. The hardware used in o…
Figure 6
Figure 6. Figure 6: Trajectory Comparison on the RE10K [46]. Our method and the baseline are under 16-view input. Our method achieves higher pose estimation accuracy than other unposed methods and is applicable to various scenes and camera motions. estimation. As shown in Tab. 3 and [PIT…
Figure 7
Figure 7. Figure 7: Additional quantitative comparison on RE10K showing runtime and memory usage across different input view counts. PSNR: 29.95 GT [PITH_FULL_IMAGE:figures/full_fig_p013_7.png]
Figure 8
Figure 8. Figure 8: Generalization results using a real video. Four frames [PITH_FULL_IMAGE:figures/full_fig_p013_8.png]

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

51 extracted references · 36 canonical work pages

  1. [1]

    Wasserstein barycenters can be computed in polynomial time in fixed di- mension

    Jason M Altschuler and Enric Boix-Adsera. Wasserstein barycenters can be computed in polynomial time in fixed di- mension. Journal of Machine Learning Research , 22(44): 1–19, 2021. 3

    work page 2021

  2. [2]

    Altschuler and Enric Boix-Adser `a

    Jason M. Altschuler and Enric Boix-Adser `a. Wasserstein barycenters are np-hard to compute. SIAM Journal on Math- ematics of Data Science, 4(1):179–203, 2022. 4

    work page 2022

  3. [3]

    Mip-nerf 360: Unbounded anti-aliased neural radiance fields

    Jonathan T Barron, Ben Mildenhall, Dor Verbin, Pratul P Srinivasan, and Peter Hedman. Mip-nerf 360: Unbounded anti-aliased neural radiance fields. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 5470–5479, 2022. 2

    work page 2022

  4. [4]

    Nope-nerf: Optimising neu- ral radiance field with no pose prior

    Wenjing Bian, Zirui Wang, Kejie Li, Jia-Wang Bian, and Victor Adrian Prisacariu. Nope-nerf: Optimising neu- ral radiance field with no pose prior. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 4160–4169, 2023. 2, 6, 7

    work page 2023

  5. [5]

    pixelsplat: 3d gaussian splats from image pairs for scalable generalizable 3d reconstruction

    David Charatan, Sizhe Lester Li, Andrea Tagliasacchi, and Vincent Sitzmann. pixelsplat: 3d gaussian splats from image pairs for scalable generalizable 3d reconstruction. In Pro- ceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 19457–19467, 2024. 2, 3

    work page 2024

  6. [6]

    Pgsr: Planar-based gaussian splatting for ef- ficient and high-fidelity surface reconstruction

    Danpeng Chen, Hai Li, Weicai Ye, Yifan Wang, Weijian Xie, Shangjin Zhai, Nan Wang, Haomin Liu, Hujun Bao, and Guofeng Zhang. Pgsr: Planar-based gaussian splatting for ef- ficient and high-fidelity surface reconstruction. IEEE Trans- actions on Visualization and Computer Graphics, 2024. 2

    2024

  7. [7]

    A survey on 3d gaussian splatting

    Guikun Chen and Wenguan Wang. A survey on 3d gaussian splatting. arXiv preprint arXiv:2401.03890, 2024. 1

    arXiv 2024

  8. [8]

    Dbarf: Deep bundle-adjusting generalizable neural radiance fields

    Yu Chen and Gim Hee Lee. Dbarf: Deep bundle-adjusting generalizable neural radiance fields. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 24–34, 2023. 2

    work page 2023

Show all 51 references
  1. [9]

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

    Yuedong Chen, Haofei Xu, Chuanxia Zheng, Bohan Zhuang, Marc Pollefeys, Andreas Geiger, Tat-Jen Cham, and Jianfei Cai. Mvsplat: Efficient 3d gaussian splatting from sparse multi-view images. In European Conference on Computer Vision, pages 370–386. Springer, 2024. 2, 3

    2024

  2. [10]

    Graph-guided scene recon- struction from images with 3d gaussian splatting, 2025

    Chong Cheng, Gaochao Song, Yiyang Yao, Qinzheng Zhou, Gangjian Zhang, and Hao Wang. Graph-guided scene recon- struction from images with 3d gaussian splatting, 2025. 2

    2025

  3. [11]

    Outdoor monocular slam with global scale- consistent 3d gaussian pointmaps, 2025

    Chong Cheng, Sicheng Yu, Zijian Wang, Yifan Zhou, and Hao Wang. Outdoor monocular slam with global scale- consistent 3d gaussian pointmaps, 2025. 2

    2025

  4. [12]

    Lu-nerf: Scene and pose estimation by synchronizing local unposed nerfs

    Zezhou Cheng, Carlos Esteves, Varun Jampani, Abhishek Kar, Subhransu Maji, and Ameesh Makadia. Lu-nerf: Scene and pose estimation by synchronizing local unposed nerfs. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 18312–18321, 2023. 2

    2023

  5. [13]

    Frank H. Clarke. Optimization and Nonsmooth Analysis. So- ciety for Industrial and Applied Mathematics, 1990. 1

    1990

  6. [14]

    Videolifter: Lifting videos to 3d with fast hierar- chical stereo alignment, 2025

    Wenyan Cong, Kevin Wang, Jiahui Lei, Colton Stearns, Yuanhao Cai, Dilin Wang, Rakesh Ranjan, Matt Feiszli, Leonidas Guibas, Zhangyang Wang, Weiyao Wang, and Zhi- wen Fan. Videolifter: Lifting videos to 3d with fast hierar- chical stereo alignment, 2025. 6, 7

    2025

  7. [15]

    Sinkhorn distances: lightspeed computation of optimal transport

    Marco Cuturi. Sinkhorn distances: lightspeed computation of optimal transport. InProceedings of the 27th International Conference on Neural Information Processing Systems - Vol- ume 2, page 2292–2300, Red Hook, NY , USA, 2013. Curran Associates Inc. 4, 1, 2

    2013

  8. [16]

    Gaussian splatting: 3d reconstruction and novel view synthesis, a review

    Anurag Dalal, Daniel Hagen, Kjell G Robbersmyr, and Kris- tian Muri Knausg˚ard. Gaussian splatting: 3d reconstruction and novel view synthesis, a review. IEEE Access, 2024. 1

    2024

  9. [17]

    A wasserstein-type dis- tance in the space of gaussian mixture models.SIAM Journal on Imaging Sciences, 13(2):936–970, 2020

    Julie Delon and Agn `es Desolneux. A wasserstein-type dis- tance in the space of gaussian mixture models.SIAM Journal on Imaging Sciences, 13(2):936–970, 2020. 3, 4

    2020

  10. [18]

    3d gaussian splatting as new era: A survey

    Ben Fei, Jingyi Xu, Rui Zhang, Qingyuan Zhou, Weidong Yang, and Ying He. 3d gaussian splatting as new era: A survey. IEEE Transactions on Visualization and Computer Graphics, 2024. 1

    2024

  11. [19]

    Alaya, Aur ´elie Boisbunon, Stanislas Cham- bon, Laetitia Chapel, Adrien Corenflos, Kilian Fatras, Nemo Fournier, L ´eo Gautheron, Nathalie T.H

    R ´emi Flamary, Nicolas Courty, Alexandre Gramfort, Mokhtar Z. Alaya, Aur ´elie Boisbunon, Stanislas Cham- bon, Laetitia Chapel, Adrien Corenflos, Kilian Fatras, Nemo Fournier, L ´eo Gautheron, Nathalie T.H. Gayraud, Hicham Janati, Alain Rakotomamonjy, Ievgen Redko, Antoine Ro...

    2021

  12. [20]

    Colmap-free 3d gaussian splat- ting

    Yang Fu, Sifei Liu, Amey Kulkarni, Jan Kautz, Alexei A Efros, and Xiaolong Wang. Colmap-free 3d gaussian splat- ting. In Proceedings of the IEEE/CVF Conference on Com- puter Vision and Pattern Recognition , pages 20796–20805,

  13. [21]

    Pf3plat: Pose-free feed-forward 3d gaussian splatting

    Sunghwan Hong, Jaewoo Jung, Heeseong Shin, Jisang Han, Jiaolong Yang, Chong Luo, and Seungryong Kim. Pf3plat: Pose-free feed-forward 3d gaussian splatting. arXiv preprint arXiv:2410.22128, 2024. 2

    2024 arXiv

  14. [22]

    Unifying cor- respondence pose and nerf for generalized pose-free novel view synthesis

    Sunghwan Hong, Jaewoo Jung, Heeseong Shin, Jiaolong Yang, Seungryong Kim, and Chong Luo. Unifying cor- respondence pose and nerf for generalized pose-free novel view synthesis. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 20196– 2...

    2024

  15. [23]

    3d gaussian splatting for real-time radiance field rendering

    Bernhard Kerbl, Georgios Kopanas, Thomas Leimk ¨uhler, and George Drettakis. 3d gaussian splatting for real-time radiance field rendering. ACM Trans. Graph., 42(4):139–1,

  16. [24]

    Ground- ing image matching in 3d with mast3r, 2024

    Vincent Leroy, Yohann Cabon, and J´erˆome Revaud. Ground- ing image matching in 3d with mast3r, 2024. 3, 6, 7 9

    2024

  17. [25]

    Barf: Bundle-adjusting neural radiance fields

    Chen-Hsuan Lin, Wei-Chiu Ma, Antonio Torralba, and Si- mon Lucey. Barf: Bundle-adjusting neural radiance fields. In IEEE International Conference on Computer Vision (ICCV),

  18. [26]

    Infinite na- ture: Perpetual view generation of natural scenes from a sin- gle image

    Andrew Liu, Richard Tucker, Varun Jampani, Ameesh Makadia, Noah Snavely, and Angjoo Kanazawa. Infinite na- ture: Perpetual view generation of natural scenes from a sin- gle image. In Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV), 2021. 6, 7, 8

    2021

  19. [27]

    Neural sparse voxel fields

    Lingjie Liu, Jiatao Gu, Kyaw Zaw Lin, Tat-Seng Chua, and Christian Theobalt. Neural sparse voxel fields. Advances in Neural Information Processing Systems, 33:15651–15663,

  20. [28]

    Scaffold-gs: Structured 3d gaussians for view-adaptive rendering

    Tao Lu, Mulin Yu, Linning Xu, Yuanbo Xiangli, Limin Wang, Dahua Lin, and Bo Dai. Scaffold-gs: Structured 3d gaussians for view-adaptive rendering. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 20654–20664, 2024. 2

    2024

  21. [29]

    Nerf: Representing scenes as neural radiance fields for view syn- thesis

    Ben Mildenhall, Pratul P Srinivasan, Matthew Tancik, Jonathan T Barron, Ravi Ramamoorthi, and Ren Ng. Nerf: Representing scenes as neural radiance fields for view syn- thesis. Communications of the ACM , 65(1):99–106, 2021. 2

    2021

  22. [30]

    D-nerf: Neural radiance fields for dynamic scenes

    Albert Pumarola, Enric Corona, Gerard Pons-Moll, and Francesc Moreno-Noguer. D-nerf: Neural radiance fields for dynamic scenes. In Proceedings of the IEEE/CVF con- ference on computer vision and pattern recognition , pages 10318–10327, 2021. 2

    2021

  23. [31]

    Octree-gs: Towards consistent real-time rendering with lod-structured 3d gaussians

    Kerui Ren, Lihan Jiang, Tao Lu, Mulin Yu, Linning Xu, Zhangkai Ni, and Bo Dai. Octree-gs: Towards consistent real-time rendering with lod-structured 3d gaussians. arXiv preprint arXiv:2403.17898, 2024. 2

    2024 arXiv

  24. [32]

    Optimal transport for applied math- ematicians

    Filippo Santambrogio. Optimal transport for applied math- ematicians. Springer, 2015. 1

    2015

  25. [33]

    Structure-from-motion revisited

    Johannes Lutz Sch ¨onberger and Jan-Michael Frahm. Structure-from-motion revisited. In Conference on Com- puter Vision and Pattern Recognition (CVPR), 2016. 6, 7

    2016

  26. [34]

    Pixelwise view selection for un- structured multi-view stereo

    Johannes Lutz Sch ¨onberger, Enliang Zheng, Marc Pollefeys, and Jan-Michael Frahm. Pixelwise view selection for un- structured multi-view stereo. In European Conference on Computer Vision (ECCV), 2016. 6

    2016

  27. [35]

    Splatt3r: Zero-shot gaussian splatting from uncalibrated image pairs, 2024

    Brandon Smart, Chuanxia Zheng, Iro Laina, and Vic- tor Adrian Prisacariu. Splatt3r: Zero-shot gaussian splatting from uncalibrated image pairs, 2024. 2, 6, 7

    2024

  28. [36]

    Flowcam: Training generalizable 3d radiance fields without camera poses via pixel-aligned scene flow

    Cameron Smith, Yilun Du, Ayush Tewari, and Vincent Sitz- mann. Flowcam: Training generalizable 3d radiance fields without camera poses via pixel-aligned scene flow. arXiv preprint arXiv:2306.00180, 2023. 2

    2023 arXiv

  29. [37]

    Gvkf: Gaus- sian voxel kernel functions for highly efficient surface recon- struction in open scenes, 2024

    Gaochao Song, Chong Cheng, and Hao Wang. Gvkf: Gaus- sian voxel kernel functions for highly efficient surface recon- struction in open scenes, 2024. 1

    2024

  30. [38]

    Block-nerf: Scalable large scene neural view synthesis

    Matthew Tancik, Vincent Casser, Xinchen Yan, Sabeek Prad- han, Ben Mildenhall, Pratul P Srinivasan, Jonathan T Barron, and Henrik Kretzschmar. Block-nerf: Scalable large scene neural view synthesis. In Proceedings of the IEEE/CVF con- ference on computer vision and pattern rec...

    2022

  31. [39]

    Sparf: Neural radiance fields from sparse and noisy poses, 2023

    Prune Truong, Marie-Julie Rakotosaona, Fabian Manhardt, and Federico Tombari. Sparf: Neural radiance fields from sparse and noisy poses, 2023. 2

    2023

  32. [40]

    Dust3r: Geometric 3d vi- sion made easy

    Shuzhe Wang, Vincent Leroy, Yohann Cabon, Boris Chidlovskii, and Jerome Revaud. Dust3r: Geometric 3d vi- sion made easy. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 20697–20709, 2024. 3

    2024

  33. [41]

    Recent advances in 3d gaussian splatting

    Tong Wu, Yu-Jie Yuan, Ling-Xiao Zhang, Jie Yang, Yan- Pei Cao, Ling-Qi Yan, and Lin Gao. Recent advances in 3d gaussian splatting. Computational Visual Media, 10(4):613– 642, 2024. 1

    2024

  34. [42]

    Depthsplat: Connecting gaussian splatting and depth

    Haofei Xu, Songyou Peng, Fangjinhua Wang, Hermann Blum, Daniel Barath, Andreas Geiger, and Marc Pollefeys. Depthsplat: Connecting gaussian splatting and depth. arXiv preprint arXiv:2410.13862, 2024. 2

    2024 arXiv

  35. [43]

    No pose, no problem: Surprisingly simple 3d gaussian splats from sparse unposed images

    Botao Ye, Sifei Liu, Haofei Xu, Xueting Li, Marc Pollefeys, Ming-Hsuan Yang, and Songyou Peng. No pose, no problem: Surprisingly simple 3d gaussian splats from sparse unposed images. arXiv preprint arXiv:2410.24207, 2024. 2, 3, 6, 7

    2024 arXiv

  36. [44]

    Rgb-only gaussian splatting slam for unbounded outdoor scenes, 2025

    Sicheng Yu, Chong Cheng, Yifan Zhou, Xiaojun Yang, and Hao Wang. Rgb-only gaussian splatting slam for unbounded outdoor scenes, 2025. 2

    2025

  37. [45]

    Nerf++: Analyzing and improving neural radiance fields

    Kai Zhang, Gernot Riegler, Noah Snavely, and Vladlen Koltun. Nerf++: Analyzing and improving neural radiance fields. arXiv preprint arXiv:2010.07492, 2020. 2

    2010 arXiv

  38. [46]

    Stereo magnification: Learning view syn- thesis using multiplane images

    Tinghui Zhou, Richard Tucker, John Flynn, Graham Fyffe, and Noah Snavely. Stereo magnification: Learning view syn- thesis using multiplane images. In SIGGRAPH, 2018. 6, 7, 8

    2018

  39. [47]

    Long-lrm: Long- sequence large reconstruction model for wide-coverage gaussian splats, 2024

    Chen Ziwen, Hao Tan, Kai Zhang, Sai Bi, Fujun Luan, Yi- cong Hong, Li Fuxin, and Zexiang Xu. Long-lrm: Long- sequence large reconstruction model for wide-coverage gaussian splats, 2024. 2 10 RegGS: Unposed Sparse Views Gaussian Splatting with 3DGS Registration Supplementary Material

    2024

  40. [48]

    As ϵ → 0, the gradient ∇ξW 2 2,ϵ converges to the subgradient set of the exact Wasserstein distance, denoted as ∂W 2 2 (ξ): lim ϵ→0 ∇ξW 2 2,ϵ ∈ ∂W 2 2 (ξ)

    Entropy-Regularized SinkhornW 2 Distance Gradient Consistency Proof The entropy-regularized Wasserstein distance W 2 2,ϵ(GA, T(GB)), where ϵ > 0 is a regularization parameter, provides a computationally feasible approach to the infinite-dimensional optimization problem inheren...

  41. [49]

    Sinkhorn Algorithm Complexity The main map and the submap contain M and N Gaus- sian gradients, respectively. Initially, the first step of the Sinkhorn algorithm involves constructing a kernel matrix 1 K ∈ RM ×N , whose elements are given by Kik = exp − Cik ϵ , (26) where Cik ...

  42. [50]

    Additional quantitative comparison on RE10K showing runtime and memory usage across different input view counts

    Additional Experimental Results Figure 8 illustrates the generalization of our method on real video data, where we uniformly sampled four frames from 2 Method 2× 16× 64× PSNR↑ Time↓ GPU(GB)↓ PSNR↑ Time↓ GPU(GB)↓ PSNR↑ Time↓ GPU(GB)↓ Splatt3R 13.951 20s 7.3 - - - - - - NoPoSpla...

  43. [51]

    Additional Limitations Figure 9 demonstrates that NoPoSplat generates suboptimal Gaussians in certain scenes. In the depicted scenario, No- PoSplat struggles to accurately estimate the depth informa- tion of the reflective surface, causing the gaussians to fail at capturing th...

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.