Pith. sign in

REVIEW 3 major objections 4 minor 1 cited by

Improving Geometry in Sparse-View 3DGS via Reprojection-based DoF Separation

T0 review · 3 major / 4 minor · reviewed 2026-08-11 · deepseek-v4-flash

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

Pith's one-line read Separating a Gaussian's side-to-side and depth motions keeps sparse-view 3D geometry from bending into texture.

desk verdict Useful technique for improving geometry in sparse-view 3DGS, but the geometry claim rests on a monocular depth-correlation proxy that needs ground-truth validation. read the letter →

arxiv 2412.14568 v1 pith:TVKAMK5F submitted 2024-12-19 cs.CV

Yongsung Kim , Minjun Park , Jooyoung Choi , Sungroh Yoon This is my paper · ORCID
classification cs.CV
keywords sparse-view3DreconstructionGaussiansplattingdegree-of-freedomseparationvisibilitylossboundedoffsetpatch-wisedepthcorrelationgeometricplausibilitymulti-viewstereorefinement
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 argues that the geometric artifacts seen when refining sparse-view reconstructions with 3D Gaussian splatting come from giving each Gaussian too much positional freedom. The authors split the three positional degrees of freedom into two kinds with different uncertainty: motion parallel to the image plane, which is tightly constrained by pixel observations, and motion along the camera ray, which is ambiguous from any single view. They manage these separately, limiting in-plane motion to a half-pixel offset and using a depth-based visibility loss to refine ray-aligned depth. The result, they report across three benchmarks, is reconstruction with both higher geometric plausibility and maintained or improved rendering quality.

What carries the argument

The carrying object is the reprojection-based DoF separation, which recovers each Gaussian from a pixel coordinate, a depth value, and a learned 2D offset rather than optimizing an unconstrained 3D position. The image-plane-parallel DoFs are controlled by the bounded offset of Eq. (2), keeping the point inside the pixel's viewing frustum, and the ray-aligned DoF is controlled by the visibility loss of Eq. (4), which penalizes disagreement between the rendered depth and the per-view depth and thereby resolves occlusion conflicts when multiple views are fused.

What would settle it

Compare the method's rendered depth maps against ground-truth 3D scans (e.g., the laser-scanned point clouds publicly available for the Tanks and Temples benchmark). If the large PDC gains over baselines do not translate into lower mean absolute depth error or higher inlier ratios relative to those scans, the central claim of improved geometry would be falsified.

Watch

Extended reading notes

Core claim

The central discovery is that separating positional degrees of freedom by uncertainty and constraining them accordingly — a bounded image-plane offset $\delta_i = 0.5\tanh(o_i)$ and a visibility loss $L_{\text{vis}} = \|\hat{D}_i - D_i\|_2^2$ comparing rendered depth to per-view MVS depth — suppresses texture-driven geometric distortions such as bumps, gaps, and floaters. Starting from a learning-based multi-view stereo point cloud with per-view depth estimates, the method unprojects pixels into 3D Gaussians and re-renders, so the two DoF types are explicitly decoupled. Experiments on Mip-NeRF 360, MVImgNet, and Tanks and Temples show consistent improvements in patch-wise depth Pearson correlation against monocular depth, with PSNR, SSIM, and LPIPS comparable to or better than strong baselines.

Load-bearing premise

The weakest load-bearing premise is that patch-wise Pearson correlation between rendered depth and a monocular depth estimator's predictions is a valid measure of true geometric plausibility; if high correlation merely reflects agreement with a learned prior rather than accurate scene geometry, the reported gains would not establish geometric fidelity.

Editorial extensions

If this is right

  • Sparse-view 3D Gaussian refinement can be made geometrically stable without sacrificing rendering quality, as shown by the PSNR, SSIM, and LPIPS results in Tables 1 and 2.
  • The method works even in extremely sparse settings: training on only three views from Tanks and Temples lifts the patch-wise depth correlation from 0.3006 for the strongest baseline to 0.6019.
  • Per-view depth maps from the MVS model are retained as a trainable target, so the intermediate geometry prior is used during refinement instead of being discarded.
  • The separation principle gives a generic recipe: constrain tightly the parameters that are well-determined by the data and use multi-view evidence for the underdetermined ones.

Reading between the lines

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

  • A direct extension the paper does not test: the same DoF separation could be applied to other scene representations (e.g., neural radiance field samples or mesh vertices) whenever a per-ray depth prior exists.
  • Because the reported PDC metric compares against a learned monocular depth prior, validating against ground-truth scans (available in Tanks and Temples) would reveal whether the large PDC gains reflect genuine metric accuracy or mere alignment with that prior.
  • The bounded offset behaves like an implicit regularization that keeps each Gaussian on the 'surface of visibility' of the input views; an adaptive bound based on per-pixel uncertainty could be a natural follow-up.
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

3 major / 4 minor

Summary. The paper addresses geometric artifacts that arise when 3D Gaussian Splatting (3DGS) is used to refine point clouds produced by learning-based multi-view stereo (MVS) models such as MASt3R in sparse-view settings. The authors propose reprojection-based DoF separation, splitting the three positional degrees of freedom of Gaussians into two image-plane-parallel DoFs with a bounded offset of at most half a pixel, and one ray-aligned DoF supervised by a visibility loss against the per-view depths from MASt3R. Experiments on Mip-NeRF 360, MVImgNet, and Tanks and Temples report improved patch-wise depth Pearson correlation (PDC) while maintaining rendering quality, plus ablations for the two proposed constraints.

Significance. If the geometry improvement is substantiated, the method offers a practical and conceptually clean way to refine MVS-initialized 3DGS without sacrificing rendering quality. The two constraints are simple to implement, and the paper explicitly identifies a plausible cause of geometric overfitting—excessive positional DoFs. However, the central claim of improved geometry is quantified almost exclusively through PDC against Depth Anything V2, a learned monocular depth estimator, and the method itself pulls geometry toward MASt3R's per-view depths, which are also learned predictions. Without ground-truth geometric evaluation or control experiments separating the effect of the depth loss from the DoF separation, the evidence for actual geometric accuracy is not yet established. The rendering-quality results are trustworthy and the ablations are informative, but the key evaluation metric needs stronger validation.

major comments (3)
  1. [Sec. 5.1 (Metrics); Tables 1-3] The central geometry claim is supported only by patch-wise depth Pearson correlation (PDC) against Depth Anything V2 monocular depth estimates. Because the visibility loss in Eq. (4) explicitly supervises the ray-aligned DoF with the per-view depths Di produced by MASt3R, and MASt3R is itself a learned MVS model, the high PDC values may reflect agreement between two learned depth priors rather than accuracy relative to true surfaces. Please add a quantitative comparison against ground-truth geometry, e.g., using the Tanks and Temples laser scans with Chamfer distance, F-score, or absolute depth error, and report a scale-sensitive metric or scale-invariant depth error in addition to Pearson correlation.
  2. [Tables 1-3] No error bars or per-scene standard deviations are reported, so the significance of the PDC gaps (e.g., 0.6019 vs. 0.2452 in Table 2) is unclear. Per-scene results in Table 3 show small differences between Ours and w/o Offset (e.g., Ballroom 0.8114 vs. 0.8093 at 12 shots), so the bounded offset's contribution to geometry is marginal; please report variance across scenes and discuss the per-scene variability, or provide a statistical test.
  3. [Sec. 4.3, Table 3] The visibility loss appears to be the primary driver of the PDC gains, but the paper does not compare against a standard 3DGS baseline that uses the same depth-supervision loss without the proposed DoF separation. Such a control would isolate whether the improvement stems from the depth loss itself or from the separation of positional DoFs. Please add this ablation.
minor comments (4)
  1. [Figure 1 caption] The abbreviation PDC is used in the figure caption but introduced only later in Sec. 5.1; please define it at first use in the figure caption or move the definition earlier.
  2. [Reference [19]] The reference list entry for SparseNeRF omits the first author's full name; it should read 'Guangcong Li, Zhaoxi Chen, Chen Change Loy, and Ziwei Liu' rather than 'Guangcong, Zhaoxi Chen, ...'.
  3. [Sec. 5.2 (Quantitative results)] Please clarify the number of test views and whether they are held-out views not used for training, as this affects interpretation of the reported PSNR and PDC scores.
  4. [Eq. (1)] The notation Ti is used for the transformation from pixel to world coordinates, but in the equation it could be misread as a tensor of all camera parameters; please clarify that Ti represents per-view camera pose and intrinsics, and define the output point cloud Xi consistently.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the geometry metric (PDC against Depth Anything V2) is external to the training objective, and the MASt3R depth prior is an explicit inductive bias rather than a hidden fit.

full rationale

The derivation chain is not circular. The geometric output is parameterized by unprojecting per-view MVS depths (Eq. 3), and Eq. 4 (Lvis = ||Dhat_i - D_i||^2) deliberately anchors rendered depth to the MASt3R per-view depth used for initialization. This is an explicit inductive bias, not a hidden fit: the paper states that retaining per-view depth is a 'secondary benefit' and that the visibility loss preserves 'valuable per-view depth information from the MVS model.' The central geometry claim is not, however, evaluated against that same Di: the PDC metric compares rendered depth maps to Depth Anything V2 monocular estimates (Sec. 5.1), an external predictor not used in the training loss. Because the reported geometry metric is not the minimized quantity, the high PDC values are not forced by construction. Nor does the paper invoke a self-citation chain or uniqueness theorem to forbid alternative approaches; MASt3R and DUSt3R are external prior works, and the method's components (tanh-bounded offsets, visibility loss) are introduced in this paper. The absence of ground-truth depth evaluation on Tanks and Temples is a genuine limitation in the strength of the 'geometrically plausible' claim, but that is an evidentiary validity concern, not a circularity: the paper's equations do not reduce to its evaluation metric. The limitation section additionally acknowledges failures inherited from MASt3R, confirming that the MVS model is treated as an external component rather than as a self-justifying oracle. Overall, no equation-level tautology, no fitted parameter renamed as prediction, and no load-bearing self-citation chain are present.

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

The method introduces no new physical entities and makes no first-principles derivation. It relies on MASt3R's per-view depths as geometric targets, a monocular depth estimator as evaluation proxy, and several hand-chosen hyperparameters; the visibility loss weight is not reported. The empirical claim is therefore dependent on the quality of external depth priors.

free parameters (4)
  • offset bound (0.5 pixel) = 0.5
    Maximum image-plane offset in Eq. (2); chosen by hand to keep unprojected point inside the pixel frustum; not optimized or justified by data.
  • visibility loss weight schedule = not specified in text
    Paper states linear scheduling decreasing over 10,000 iterations (Sec. 5.1) but never gives the initial or final weight; the central regularization strength is therefore unreported.
  • scale clipping threshold = 30 pixels
    Gaussian scale clipped to 30 pixels following MASt3R's configuration (Sec. 5.1); a hand-picked hyperparameter affecting scene coverage.
  • number of training iterations = 10000
    Training budget chosen for all experiments (Sec. 5.1); typical hyperparameter, not fitted to data.
assumptions (5)
  • domain assumption Image-plane-parallel DoFs have low uncertainty bounded by pixel size, while the ray-aligned DoF has high uncertainty requiring multi-view integration.
    Core premise of Sec. 4.2 motivating the whole split; not proven, just asserted.
  • domain assumption MASt3R's per-view depth maps after global alignment are ambiguity-free and accurate enough to supervise ray-aligned refinement.
    Sec. 4.2 and 4.3; the visibility loss Eq. (4) directly pulls rendered depth toward these depths. If the MVS depths are wrong, the error is baked in; paper acknowledges inherited failures in Sec. 6.
  • domain assumption Rendered depth smaller than per-view depth implies occlusion by another Gaussian.
    Sec. 4.3 visibility-loss motivation; a simplifying heuristic that ignores transparency and semi-transparent Gaussians.
  • domain assumption Monocular depth estimator is a valid reference for geometric plausibility.
    Sec. 5.1 metrics; PDC uses Depth Anything V2 estimates instead of ground-truth geometry.
  • standard math Unprojecting within a 0.5-pixel offset guarantees the Gaussian projects back onto its original pixel under the pinhole model.
    Used to justify bounded offset in Sec. 4.3; true for ideal pinhole geometry and no distortion, but real lenses or datasets may deviate.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Improving Geometry in Sparse-View 3DGS via Reprojection-based DoF Separation." pith.science (2026). https://pith.science/paper/TVKAMK5F

@misc{pith2026241214568,
  author       = {Pith},
  title        = {Pith review of: Improving Geometry in Sparse-View 3DGS via Reprojection-based DoF Separation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/TVKAMK5F}},
  note         = {Machine review of arXiv:2412.14568}
}
read the original abstract

Recent learning-based Multi-View Stereo models have demonstrated state-of-the-art performance in sparse-view 3D reconstruction. However, directly applying 3D Gaussian Splatting (3DGS) as a refinement step following these models presents challenges. We hypothesize that the excessive positional degrees of freedom (DoFs) in Gaussians induce geometry distortion, fitting color patterns at the cost of structural fidelity. To address this, we propose reprojection-based DoF separation, a method distinguishing positional DoFs in terms of uncertainty: image-plane-parallel DoFs and ray-aligned DoF. To independently manage each DoF, we introduce a reprojection process along with tailored constraints for each DoF. Through experiments across various datasets, we confirm that separating the positional DoFs of Gaussians and applying targeted constraints effectively suppresses geometric artifacts, producing reconstruction results that are both visually and geometrically plausible.

Figures

Figures reproduced from arXiv: 2412.14568 by the authors.

Figure 1
Figure 1. 3D reconstruction results with 12 views. The image on the left side is the rendered depth image, and the upper right image is the rendered RGB image. The lower right image visualizes patch-wise depth correlation, where green indicates accurate geometry. As the color shifts from green to gray and then to purple, the patch-wise depth correlation decreases, indicating less plausible geometry. Our method qualitatively d… view at source ↗
Figure 2
Figure 2. Geometric artifacts from naive 3DGS refinement. Texture representation via Gaussian positions introduces unin￾tended geometric patterns. The red box highlights excessive dis￾tortion in the ceiling geometry, and the blue box shows gaps in the flat floor geometry following texture patterns. learning frameworks, it can lead to suboptimal geometry be￾cause the optimization process disregards information from other views… view at source ↗
Figure 3
Figure 3. Overview of the proposed framework. (a) Scene initialization using a learning-based MVS model, which predicts 3D points from images and outputs per-view depth as an intermediate representation. (b) Na¨ıve implementation, where MVS is treated as a black-box model, and its output is refined using the 3DGS pipeline. (c) Our proposed framework, which introduces reprojection-based refinement by retaining intermediate per… view at source ↗
Figures from the paper (5 more)
Figure 4
Figure 4. Figure 4: Separation of DoFs. The left column illustrates the two types of positional degrees of freedom (DoFs): ray-aligned and image-plane-parallel. The right column shows two scenarios where a point fails to project back to its original pixel. Bounded Offset. For the first co…
Figure 5
Figure 5. Figure 5: Qualitative comparison on Tanks and Temples. Our method outperforms the baselines not only in reconstructing smooth surfaces, such as the floor, but also in capturing the geometry of complex shapes like statues. As the color shifts from green to gray and then to purple…
Figure 6
Figure 6. Figure 6: Qualitative comparison on Mip-NeRF 360. In the Counter scene, the baselines represent the patterns of the tray and tablecloth as geometric artifacts, whereas our method more plausibly captures the geometry. Additionally, in the Bonsai scene, the baselines produce numer…
Figure 7
Figure 7. Figure 7: Qualitative comparison on MVImgNet. The baselines distort the geometry to represent the texture of the surface of the chair and ground, resulting in geometric artifacts that are identifiable not only in the depth map but also in the RGB image. In the SUV scene, our met…
Figure 9
Figure 9. Figure 9: Limitations inherited from the MASt3R. The first row illustrates failures from MASt3R’s inaccurate camera pose estima￾tion, resulting in the bicycle wheel’s overlapping geometries. The second row presents failures caused by specular surfaces, where the geometry of the …

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. Sparse-View 3D Reconstruction: Recent Advances and Open Challenges

    cs.CV 2025-07 conditional novelty 3.0 of 10

    A comprehensive survey that organizes sparse-view 3D reconstruction methods into geometry-based, NeRF, 3DGS, and diffusion-based categories, with benchmarks and open challenges.

Reference graph

Works this paper leans on

62 extracted references · 43 canonical work pages · cited by 1 Pith paper

  1. [1]

    Amodal 3d reconstruction for robotic manipulation via stability and connectivity

    William Agnew, Christopher Xie, Aaron Walsman, Octavian Murad, Yubo Wang, Pedro Domingos, and Siddhartha Srini- vasa. Amodal 3d reconstruction for robotic manipulation via stability and connectivity. In Conference on Robot Learning, pages 1498–1508. PMLR, 2021. 1

    work page 2021

  2. [2]

    Barron, Ben Mildenhall, Dor Verbin, Pratul P

    Jonathan T. Barron, Ben Mildenhall, Dor Verbin, Pratul P. Srinivasan, and Peter Hedman. Mip-nerf 360: Unbounded anti-aliased neural radiance fields. CVPR, 2022. 2, 5, 6

    work page 2022

  3. [3]

    Surf: Speeded up robust features

    Herbert Bay. Surf: Speeded up robust features. Computer Vision—ECCV, 2006. 1, 2

    work page 2006

  4. [4]

    Nope-nerf: Optimising neural radiance field with no pose prior

    Wenjing Bian, Zirui Wang, Kejie Li, Jiawang Bian, and Vic- tor Adrian Prisacariu. Nope-nerf: Optimising neural radiance field with no pose prior. In CVPR, 2023. 5, 6

    work page 2023

  5. [5]

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

    David Charatan, Sizhe Li, Andrea Tagliasacchi, and Vincent Sitzmann. pixelsplat: 3d gaussian splats from image pairs for scalable generalizable 3d reconstruction. In CVPR, 2024. 2

    work page 2024

  6. [6]

    3d point cloud processing and learning for autonomous driving: Impacting map cre- ation, localization, and perception

    Siheng Chen, Baoan Liu, Chen Feng, Carlos Vallespi- Gonzalez, and Carl Wellington. 3d point cloud processing and learning for autonomous driving: Impacting map cre- ation, localization, and perception. IEEE Signal Processing Magazine, 38(1):68–86, 2021. 1

    work page 2021

  7. [7]

    Deep stereo using adap- tive thin volume representation with uncertainty awareness

    Shuo Cheng, Zexiang Xu, Shilin Zhu, Zhuwen Li, Li Erran Li, Ravi Ramamoorthi, and Hao Su. Deep stereo using adap- tive thin volume representation with uncertainty awareness. In Proceedings of the IEEE/CVF conference on computer vi- sion and pattern recognition, pages 2524–2534, 2020. 2

    work page 2020

  8. [8]

    3d-r2n2: A unified approach for single and multi-view 3d object reconstruction

    Christopher B Choy, Danfei Xu, JunYoung Gwak, Kevin Chen, and Silvio Savarese. 3d-r2n2: A unified approach for single and multi-view 3d object reconstruction. In Computer Vision–ECCV 2016: 14th European Conference, Amsterdam, The Netherlands, October 11-14, 2016, Proceedings, Part VIII 14, pages 628–644. Springer, 2016. 2

    work page 2016

Show all 62 references
  1. [9]

    Depth-supervised nerf: Fewer views and faster train- ing for free

    Kangle Deng, Andrew Liu, Jun-Yan Zhu, and Deva Ra- manan. Depth-supervised nerf: Fewer views and faster train- ing for free. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 12882– 12891, 2022. 2, 3

    2022

  2. [10]

    An image is worth 16x16 words: Transformers for image recognition at scale

    Alexey Dosovitskiy, Lucas Beyer, Alexander Kolesnikov, Dirk Weissenborn, Xiaohua Zhai, Thomas Unterthiner, Mostafa Dehghani, Matthias Minderer, Georg Heigold, Syl- vain Gelly, Jakob Uszkoreit, and Neil Houlsby. An image is worth 16x16 words: Transformers for image recognition ...

    2021

  3. [11]

    Lsd- slam: Large-scale direct monocular slam

    Jakob Engel, Thomas Sch ¨ops, and Daniel Cremers. Lsd- slam: Large-scale direct monocular slam. In European con- ference on computer vision, pages 834–849. Springer, 2014. 2

    2014

  4. [12]

    Instantsplat: Unbounded sparse-view pose-free gaus- sian splatting in 40 seconds, 2024

    Zhiwen Fan, Wenyan Cong, Kairun Wen, Kevin Wang, Jian Zhang, Xinghao Ding, Danfei Xu, Boris Ivanovic, Marco Pavone, Georgios Pavlakos, Zhangyang Wang, and Yue Wang. Instantsplat: Unbounded sparse-view pose-free gaus- sian splatting in 40 seconds, 2024. 3, 5, 6

    2024

  5. [13]

    Fusion of multiview silhou- ette cues using a space occupancy grid

    J-S Franco and Edmond Boyer. Fusion of multiview silhou- ette cues using a space occupancy grid. In Tenth IEEE Inter- national Conference on Computer Vision (ICCV’05) Volume 1, pages 1747–1753. IEEE, 2005. 2

    2005

  6. [14]

    Mve-a multi-view reconstruction environment

    Simon Fuhrmann, Fabian Langguth, and Michael Goesele. Mve-a multi-view reconstruction environment. GCH, 3:4,

  7. [15]

    Accurate, dense, and robust multiview stereopsis

    Yasutaka Furukawa and Jean Ponce. Accurate, dense, and robust multiview stereopsis. IEEE transactions on pattern analysis and machine intelligence , 32(8):1362–1376, 2009. 1

    2009

  8. [16]

    Towards internet-scale multi-view stereo

    Yasutaka Furukawa, Brian Curless, Steven M Seitz, and Richard Szeliski. Towards internet-scale multi-view stereo. In 2010 IEEE computer society conference on computer vi- sion and pattern recognition, pages 1434–1441. IEEE, 2010. 2

    2010

  9. [17]

    Massively parallel multiview stereopsis by surface normal diffusion

    Silvano Galliani, Katrin Lasinger, and Konrad Schindler. Massively parallel multiview stereopsis by surface normal diffusion. In Proceedings of the IEEE international confer- ence on computer vision, pages 873–881, 2015. 2

    2015

  10. [18]

    Cascade cost volume for high-resolution multi-view stereo and stereo matching

    Xiaodong Gu, Zhiwen Fan, Siyu Zhu, Zuozhuo Dai, Feitong Tan, and Ping Tan. Cascade cost volume for high-resolution multi-view stereo and stereo matching. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 2495–2504, 2020. 2

    2020

  11. [19]

    Sparsenerf: Distilling depth ranking for few-shot novel view synthesis

    Guangcong, Zhaoxi Chen, Chen Change Loy, and Ziwei Liu. Sparsenerf: Distilling depth ranking for few-shot novel view synthesis. IEEE/CVF International Conference on Computer Vision (ICCV), 2023. 2, 3

    2023

  12. [20]

    Putting nerf on a diet: Semantically consistent few-shot view synthesis

    Ajay Jain, Matthew Tancik, and Pieter Abbeel. Putting nerf on a diet: Semantically consistent few-shot view synthesis. In ICCV, pages 5885–5894, 2021. 2, 3

    2021

  13. [21]

    Surfacenet: An end-to-end 3d neural network for mul- tiview stereopsis

    Mengqi Ji, Juergen Gall, Haitian Zheng, Yebin Liu, and Lu Fang. Surfacenet: An end-to-end 3d neural network for mul- tiview stereopsis. In Proceedings of the IEEE international conference on computer vision, pages 2307–2315, 2017. 2

    2017

  14. [22]

    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,

  15. [23]

    Tanks and temples: Benchmarking large-scale scene reconstruction

    Arno Knapitsch, Jaesik Park, Qian-Yi Zhou, and Vladlen Koltun. Tanks and temples: Benchmarking large-scale scene reconstruction. ACM Transactions on Graphics (ToG) , 36 (4):1–13, 2017. 2, 5

    2017

  16. [24]

    A theory of shape by space carving

    Kiriakos N Kutulakos and Steven M Seitz. A theory of shape by space carving. International journal of computer vision , 38:199–218, 2000. 2 9

    2000

  17. [25]

    Gecon- erf: Few-shot neural radiance fields via geometric consis- tency

    Min-Seop Kwak, Jiuhn Song, and Seungryong Kim. Gecon- erf: Few-shot neural radiance fields via geometric consis- tency. In ICML, pages 18023–18036. PMLR, 2023. 2, 3

    2023

  18. [26]

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

    Vincent Leroy, Yohann Cabon, and Jerome Revaud. Ground- ing image matching in 3d with mast3r, 2024. 1, 2, 3, 6

    2024

  19. [27]

    Dngaussian: Optimizing sparse-view 3d gaussian radiance fields with global-local depth normaliza- tion

    Jiahe Li, Jiawei Zhang, Xiao Bai, Jin Zheng, Xin Ning, Jun Zhou, and Lin Gu. Dngaussian: Optimizing sparse-view 3d gaussian radiance fields with global-local depth normaliza- tion. In CVPR, 2024. 2, 3, 5

    2024

  20. [28]

    Object recognition from local scale-invariant features

    David G Lowe. Object recognition from local scale-invariant features. In Proceedings of the seventh IEEE interna- tional conference on computer vision , pages 1150–1157. Ieee, 1999. 1, 2

    1999

  21. [29]

    Srinivasan, Matthew Tancik, Jonathan T

    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. In ECCV, 2020. 1, 2

    2020

  22. [30]

    Orb-slam: a versatile and accurate monocular slam system

    Raul Mur-Artal, Jose Maria Martinez Montiel, and Juan D Tardos. Orb-slam: a versatile and accurate monocular slam system. IEEE transactions on robotics , 31(5):1147–1163,

  23. [31]

    Atlas: End- to-end 3d scene reconstruction from posed images

    Zak Murez, Tarrence Van As, James Bartolozzi, Ayan Sinha, Vijay Badrinarayanan, and Andrew Rabinovich. Atlas: End- to-end 3d scene reconstruction from posed images. In Com- puter Vision–ECCV 2020: 16th European Conference, Glas- gow, UK, August 23–28, 2020, Proceedings, Part V...

    2020

  24. [32]

    Kinectfusion: Real-time dense surface mapping and track- ing

    Richard A Newcombe, Shahram Izadi, Otmar Hilliges, David Molyneaux, David Kim, Andrew J Davison, Pushmeet Kohi, Jamie Shotton, Steve Hodges, and Andrew Fitzgibbon. Kinectfusion: Real-time dense surface mapping and track- ing. In 2011 10th IEEE international symposium on mixed ...

    2011

  25. [33]

    Barron, Ben Mildenhall, Mehdi S

    Michael Niemeyer, Jonathan T. Barron, Ben Mildenhall, Mehdi S. M. Sajjadi, Andreas Geiger, and Noha Radwan. Regnerf: Regularizing neural radiance fields for view syn- thesis from sparse inputs. In CVPR, 2022. 2

    2022

  26. [34]

    Real-time 3d reconstruction at scale us- ing voxel hashing

    Matthias Nießner, Michael Zollh ¨ofer, Shahram Izadi, and Marc Stamminger. Real-time 3d reconstruction at scale us- ing voxel hashing. ACM Transactions on Graphics (ToG), 32 (6):1–11, 2013. 1

    2013

  27. [35]

    Beyondpixels: A com- prehensive review of the evolution of neural radiance fields

    AKM Rabby and Chengcui Zhang. Beyondpixels: A com- prehensive review of the evolution of neural radiance fields. arXiv preprint arXiv:2306.03000, 2023. 2

    2023 arXiv

  28. [36]

    Towards robust monocular depth estimation: Mixing datasets for zero-shot cross-dataset transfer

    Ren ´e Ranftl, Katrin Lasinger, David Hafner, Konrad Schindler, and Vladlen Koltun. Towards robust monocular depth estimation: Mixing datasets for zero-shot cross-dataset transfer. IEEE Transactions on Pattern Analysis and Ma- chine Intelligence, 44(3), 2022. 4

    2022

  29. [37]

    Structure- from-motion revisited

    Johannes L Schonberger and Jan-Michael Frahm. Structure- from-motion revisited. In CVPR, pages 4104–4113, 2016. 1, 3, 5

    2016

  30. [38]

    Structure- from-motion revisited

    Johannes L Schonberger and Jan-Michael Frahm. Structure- from-motion revisited. In Proceedings of the IEEE con- ference on computer vision and pattern recognition , pages 4104–4113, 2016. 2

    2016

  31. [39]

    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. 1

    2016

  32. [40]

    Photorealistic scene reconstruction by voxel coloring

    Steven M Seitz and Charles R Dyer. Photorealistic scene reconstruction by voxel coloring. International journal of computer vision, 35:151–173, 1999. 2

    1999

  33. [41]

    Seitz, and Richard Szeliski

    Noah Snavely, Steven M. Seitz, and Richard Szeliski. Photo tourism: Exploring photo collections in 3d. In SIGGRAPH Conference Proceedings, pages 835–846, New York, NY , USA, 2006. ACM Press. 2

    2006

  34. [42]

    D ¨arf: Boosting radiance fields from sparse input views with monocular depth adaptation

    Jiuhn Song, Seonghoon Park, Honggyu An, Seokju Cho, Min-Seop Kwak, Sungjin Cho, and Seungryong Kim. D ¨arf: Boosting radiance fields from sparse input views with monocular depth adaptation. Advances in Neural Informa- tion Processing Systems, 36:68458–68470, 2023. 4

    2023

  35. [43]

    Neuralrecon: Real-time coherent 3d re- construction from monocular video

    Jiaming Sun, Yiming Xie, Linghao Chen, Xiaowei Zhou, and Hujun Bao. Neuralrecon: Real-time coherent 3d re- construction from monocular video. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 15598–15607, 2021. 2

    2021

  36. [44]

    Scade: Nerfs from space carving with ambiguity-aware depth estimates

    Mikaela Angelina Uy, Ricardo Martin-Brualla, Leonidas Guibas, and Ke Li. Scade: Nerfs from space carving with ambiguity-aware depth estimates. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 16518–16527, 2023. 4

    2023

  37. [45]

    Attention is all you need

    Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszko- reit, Llion Jones, Aidan N Gomez, Ł ukasz Kaiser, and Illia Polosukhin. Attention is all you need. In Advances in Neu- ral Information Processing Systems. Curran Associates, Inc.,

  38. [46]

    Patchmatchnet: Learned multi-view patchmatch stereo

    Fangjinhua Wang, Silvano Galliani, Christoph V ogel, Pablo Speciale, and Marc Pollefeys. Patchmatchnet: Learned multi-view patchmatch stereo. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 14194–14203, 2021. 2

    2021

  39. [47]

    Learning-based multi-view stereo: A survey

    Fangjinhua Wang, Qingtian Zhu, Di Chang, Quankai Gao, Junlin Han, Tong Zhang, Richard Hartley, and Marc Polle- feys. Learning-based multi-view stereo: A survey. arXiv preprint arXiv:2408.15235, 2024. 2

    2024

  40. [48]

    Mvdepthnet: Real-time multiview depth estimation neural network

    Kaixuan Wang and Shaojie Shen. Mvdepthnet: Real-time multiview depth estimation neural network. In2018 Interna- tional conference on 3d vision (3DV), pages 248–257. IEEE,

  41. [49]

    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. In CVPR, 2024. 1, 2, 3

    2024

  42. [50]

    Image quality assessment: from error visibility to structural similarity

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

    2004

  43. [51]

    Sparsegs: Real- time 360° sparse view synthesis using gaussian splatting

    Haolin Xiong, Sairisheek Muttukuru, Rishi Upadhyay, Pradyumna Chari, and Achuta Kadambi. Sparsegs: Real- time 360° sparse view synthesis using gaussian splatting. Arxiv, 2023. 2, 3

    2023

  44. [52]

    Gaussianob- 10 ject: High-quality 3d object reconstruction from four views with gaussian splatting

    Chen Yang, Sikuang Li, Jiemin Fang, Ruofan Liang, Lingxi Xie, Xiaopeng Zhang, Wei Shen, and Qi Tian. Gaussianob- 10 ject: High-quality 3d object reconstruction from four views with gaussian splatting. In ACM TOG, 2024. 3

    2024

  45. [53]

    Freenerf: Im- proving few-shot neural rendering with free frequency regu- larization

    Jiawei Yang, Marco Pavone, and Yue Wang. Freenerf: Im- proving few-shot neural rendering with free frequency regu- larization. In CVPR, 2023. 2, 3

    2023

  46. [54]

    Depth any- thing v2

    Lihe Yang, Bingyi Kang, Zilong Huang, Zhen Zhao, Xiao- gang Xu, Jiashi Feng, and Hengshuang Zhao. Depth any- thing v2. arXiv preprint arXiv:2406.09414, 2024. 5

    2024 arXiv

  47. [55]

    Multi-resolution real- time stereo on commodity graphics hardware

    Ruigang Yang and Marc Pollefeys. Multi-resolution real- time stereo on commodity graphics hardware. In 2003 IEEE Computer Society Conference on Computer Vision and Pat- tern Recognition, 2003. Proceedings., pages I–I. IEEE, 2003. 2

    2003

  48. [56]

    Mvsnet: Depth inference for unstructured multi-view stereo

    Yao Yao, Zixin Luo, Shiwei Li, Tian Fang, and Long Quan. Mvsnet: Depth inference for unstructured multi-view stereo. In Proceedings of the European conference on computer vi- sion (ECCV), pages 767–783, 2018. 2

    2018

  49. [57]

    gsplat: An open-source library for Gaussian splatting

    Vickie Ye, Ruilong Li, Justin Kerr, Matias Turkulainen, Brent Yi, Zhuoyang Pan, Otto Seiskari, Jianbo Ye, Jeffrey Hu, Matthew Tancik, and Angjoo Kanazawa. gsplat: An open-source library for Gaussian splatting. arXiv preprint arXiv:2409.06765, 2024. 5

    2024 arXiv

  50. [58]

    Mvimgnet: A large-scale dataset of multi-view images

    Xianggang Yu, Mutian Xu, Yidan Zhang, Haolin Liu, Chongjie Ye, Yushuang Wu, Zizheng Yan, Chenming Zhu, Zhangyang Xiong, Tianyou Liang, et al. Mvimgnet: A large-scale dataset of multi-view images. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognit...

    2023

  51. [59]

    A glob- ally optimal algorithm for robust tv-l 1 range image integra- tion

    Christopher Zach, Thomas Pock, and Horst Bischof. A glob- ally optimal algorithm for robust tv-l 1 range image integra- tion. In 2007 IEEE 11th International Conference on Com- puter Vision, pages 1–8. IEEE, 2007. 2

    2007

  52. [60]

    The unreasonable effectiveness of deep features as a perceptual metric

    Richard Zhang, Phillip Isola, Alexei A Efros, Eli Shecht- man, and Oliver Wang. The unreasonable effectiveness of deep features as a perceptual metric. In Proceedings of the IEEE conference on computer vision and pattern recogni- tion, pages 586–595, 2018. 5

    2018

  53. [61]

    Patchmatch based joint view selection and depthmap estimation

    Enliang Zheng, Enrique Dunn, Vladimir Jojic, and Jan- Michael Frahm. Patchmatch based joint view selection and depthmap estimation. In Proceedings of the IEEE conference on computer vision and pattern recognition , pages 1510– 1517, 2014. 2

    2014

  54. [62]

    Fsgs: Real-time few-shot view synthesis using gaussian splatting

    Zehao Zhu, Zhiwen Fan, Yifan Jiang, and Zhangyang Wang. Fsgs: Real-time few-shot view synthesis using gaussian splatting. In ECCV, 2024. 2, 3 11

    2024

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.