Pith. sign in

REVIEW 4 major objections 6 minor 2 cited by

DET-GS: Depth- and Edge-Aware Regularization for High-Fidelity 3D Gaussian Splatting

T0 review · 4 major / 6 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 DET-GS, an extension of 3D Gaussian Splatting, claims that adding hierarchical depth supervision, Canny-edge-masked depth smoothing, and RGB-guided total variation improves sparse-view geometry and rendering quality.

desk verdict Competent incremental 3DGS paper whose headline sparse-view claim is untested — the experiments all use full training sets, so the central motivation is unsupported. read the letter →

arxiv 2508.04099 v1 pith:SDZVQHK7 submitted 2025-08-06 cs.CV cs.AI

Zexu Huang , Min Xu , Stuart Perry This is my paper
classification cs.CVcs.AI
keywords 3DGaussianSplattingnovelviewsynthesissparse-viewreconstructiondepthsupervisionmonocularestimationedge-awareregularizationtotalvariationCannyedgedetection
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

The paper proposes DET-GS, an extension of 3D Gaussian Splatting that adds three regularization losses aimed at sparse-view novel view synthesis. The central claim is that these losses—hierarchical depth supervision against monocular depth estimates, edge-aware depth smoothing masked by Canny edges, and an RGB-guided edge-preserving total variation term—together improve both the recovered geometry and the rendered images, reducing floaters and preserving sharp boundaries. The authors report consistent improvements over several prior 3DGS-based methods on five benchmark datasets, and their ablation attributes the largest gain to the hierarchical depth supervision. If the claim holds, it means noisy monocular depth can serve as a reliable geometric prior when supervision is localized and error-tolerant, and that structure-aware smoothing can replace uniform smoothing in point-based rendering.

What carries the argument

The carrying mechanism is a triple of auxiliary losses attached to the standard 3DGS optimization. The first is hierarchical depth supervision built on a re-weighted depth render (Eq. 4) that up-weights the nearest Gaussian through a parameter $\omega$, with the rendered and estimated depths compared after patch-level and image-level normalization. The second is edge-aware depth regularization: a binary mask from Canny edge detection selects non-boundary pixels, and a masked local mean over a cross-shaped kernel smooths only those pixels. The third is RGB-guided edge-preserving total variation: binary masks derived from the ground-truth RGB gradients disable the TV penalty at strong edges. T

What would settle it

Take the DET-GS pipeline and corrupt the depth prior, for example by adding noise only in texture-less regions or by applying a global scale shift, then measure PSNR/SSIM/LPIPS against the no-depth-supervision baseline; if the metrics do not drop, the claimed benefit is not coming from the depth prior. Separately, ablate the re-weighting by sweeping $\omega$ from near 0 to near 1 on a fixed Mip-NeRF 360 scene: if performance is flat across the range, the exponential depth re-weighting in Eq. 4 is not load-bearing.

Watch

Extended reading notes

Core claim

The core claim is that geometry and appearance in 3D Gaussian Splatting can both be improved under sparse views by replacing non-local depth losses and uniform smoothing with structure-aware counterparts. Depth is rendered with a modified accumulation that gives the nearest Gaussian exponentially more weight, then compared with monocular depth predictions after patch-wise and image-wise normalization, creating a dual-scale supervision that drives Gaussian centers toward consistent positions. A second loss smooths rendered depth only where the Canny edge map is zero, using a masked local mean over a four-connected neighborhood. A third loss applies total variation to the rendered image only w

Load-bearing premise

The depth-supervision branch assumes that the pretrained monocular depth predictor's maps are trustworthy in exactly the places that sparse-view reconstruction finds hardest, texture-less and occluded regions; the paper itself notes these maps can be inaccurate there, and Gaussian placement inherits those biases.

Editorial extensions

If this is right

  • Sparse-view 3DGS reconstructions will show fewer floating artifacts and sharper depth discontinuities than methods that apply uniform smoothing.
  • The hierarchical depth term, not the smoothing terms, is the main driver of the reported gains, according to the paper's ablations.
  • RGB-guided total variation smooths flat regions while preserving texture, so rendered images should be perceptually cleaner in homogeneous areas.
  • Canny-based edge masking preserves object contours that uniform depth smoothing blurs.
  • The hierarchical depth term makes the optimization more robust to depth-estimation noise than pixel-to-pixel non-local depth losses.

Reading between the lines

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

  • A natural extension is to replace the Canny edge mask with learned or confidence-weighted boundaries; the paper's own limitation section says the hand-crafted edge regularizer may not generalize across all scene types and lighting conditions.
  • The same patch/image-normalized supervision recipe could be applied to other geometry priors, such as surface normals or segmentation planes, in any point-based renderer.
  • The re-weighted depth render introduces a parameter $\omega$ that the paper does not ablate; sweeping it would reveal whether the exponential reweighting or the hierarchical normalization carries the improvement.
  • Because all three losses act entirely in image and depth space, they should transfer to any differentiable rasterizer that outputs depth, suggesting the gains are not tied to the specific 3DGS implementation.
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 / 6 minor

Summary. The manuscript proposes DET-GS, a 3D Gaussian Splatting extension with three regularization losses: (i) hierarchical depth supervision using patch-wise and image-normalized L2 losses against Depth Anything V2 predictions (Eqs. 4-7); (ii) edge-aware depth smoothing masked by Canny edges (Eqs. 8-11); and (iii) RGB-guided edge-preserving total variation (Eqs. 13-15). These are combined with a color loss in Eq. (17). Experiments compare on Mip-NeRF 360, Tanks&Temples, Deep Blending, NeRF Synthetic, and NSVF, plus ablations. The abstract claims sparse-view SOTA and geometric accuracy gains.

Significance. The paper addresses a relevant problem: using monocular depth priors to regularize 3DGS, and the three components are clearly motivated. The clean decomposition of the objective and the use of an external depth prior avoid circularity. If the sparse-view claim were demonstrated, the method would be a useful contribution to 3DGS regularization. However, as submitted the significance is conditional: no code is provided, geometric accuracy is not measured quantitatively, and the core sparse-view claim is not tested. The numerical gains on dense-view benchmarks are modest (e.g., +0.17 dB PSNR on Mip-NeRF360 vs Spec-Gaussian), which is insufficient by itself to support the stated contribution.

major comments (4)
  1. [Section IV.C (Tables I-III)] The abstract and introduction claim 'outperforming SOTA on sparse-view novel view synthesis benchmarks,' but all experiments use the full training sets of Mip-NeRF 360, Tanks&Temples, Deep Blending, NeRF Synthetic, and NSVF. There is no view subsampling (e.g., 3/6/9 views) and no sparse-view protocol such as LLFF/DTU or the DNGaussian/FSGS few-shot settings. Thus the central claim is not supported by any experiment in the paper. Please either add standard sparse-view experiments or revise the claims to dense-view benchmarks.
  2. [Section IV.C] The paper claims improvements in 'geometric accuracy' in the abstract and conclusion, but the only quantitative metrics are PSNR/SSIM/LPIPS. Depth maps are compared only qualitatively in Fig. 6. No metric such as absolute relative depth error, Chamfer distance, F-score, or comparison against a SfM/COLMAP geometry baseline is reported. Consequently the geometric-accuracy claim is not substantiated.
  3. [Section III.B, Eq. (4)] The modified depth rendering in Eq. (4) uses an undefined scalar ω and replaces the alpha-composited weights of Eq. (3) with ω exp(−ι(n−1)). There is no statement of the value, range, or schedule of ω; the phrase 'enhanced opacity values ω to all primitives' is difficult to interpret because G2D_n is already part of the projection, while the per-Gaussian opacity α_n is omitted. Since this depth is used in L_depth and L_edge, the failure to specify and justify this weighting is load-bearing. Please align with Eq. (3) or define and validate the new weighting.
  4. [Section IV.B, 'error tolerance mechanism'] The implementation section states: 'To mitigate over-constraining the learning process, we introduce an error tolerance mechanism into the depth L2 loss,' but no equation, threshold, or algorithmic description is given in Section III or Eq. (17). The loss in Eq. (7) is a plain L2 loss. This unspecified mechanism can affect all quantitative results and the ablation; without specification, the experiments are not reproducible.
minor comments (6)
  1. [Section III.B, Eq. (4)] 'where µ_i denotes the center position' should be µ_n; also the relationship between Eq. (3) and Eq. (4) is confusing because both are called D(x).
  2. [Abstract / Introduction] The phrase 'semantic masks derived from Canny edge detection' is inaccurate; Canny outputs an edge map, not semantic masks. Please reword.
  3. [Section IV.D, Table V] It is unclear what 'patch size 0' means and how 'randomized patch size sampling' is compatible with 'non-overlapping patches.' Clarify the protocol.
  4. [Table I, Deep Blending column] Scaffold-GS has PSNR 30.20 versus Ours 30.05, so the text 'consistently outperforms previous SOTA methods across all datasets' should be qualified (e.g., best on average, or per metric).
  5. [Fig. 5] The labels and numbers in the figure are hard to read; clarify which variant corresponds to each PSNR/LPIPS value and why the best scores are marked in red.
  6. [Eq. (10)] When all neighbors in the cross-shaped neighborhood are masked (m=0), the denominator reduces to epsilon and the masked mean becomes 0. Consider including a validity count to avoid numerical instability.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: all proposed losses are anchored to external priors and ground-truth images; self-citations are not load-bearing.

full rationale

The derivation chain is self-contained: each proposed loss term compares a rendered quantity to an external reference that is not derived from the model's own output or from the reported success metric. L_depth (Eq. 7) compares patch- and image-normalized rendered depth to normalized depth maps from the pretrained Depth Anything V2 model [54], an external monocular depth predictor. L_edge (Eq. 11) uses a Canny edge mask computed from the ground-truth RGB image (Eqs. 8-9) to mask a local-mean depth smoothness term; the mask and local mean come from GT structure and rendered depth, respectively, not from the target metrics. L_tv (Eq. 15) uses thresholds on ground-truth RGB gradients (Eq. 14) to gate a TV penalty on the rendered image; again the gate is external. L_color is the standard 3DGS reconstruction loss against ground-truth images. No parameter is fitted to PSNR/SSIM/LPIPS and then renamed as a prediction, and no quantity in the method is defined in terms of the claimed outcome. The only self-citations ([25], [34], [35]) appear in related-work enumerations and are not used to justify any design choice or to forbid alternatives. The paper's 'sparse-view' claim is not supported by the dense-view experimental protocol, but that is an evidence/scope issue, not a circular-derivation issue. Under-specified terms such as omega in Eq. 4 are correctness risks, not circularity.

Assumptions & free parameters 6 free parameters · 4 assumptions · 0 invented entities

The paper introduces no new physical or learned entities; it adds loss functions. The main free parameters are the unspecified omega and tuning thresholds/weights. The load-bearing assumptions concern the reliability of monocular depth priors and the alignment of RGB edges with geometric boundaries.

free parameters (6)
  • omega (depth reweighting) = not reported
    Eq. 4 weights Gaussian depth contributions by omega * exp(-iota*(n-1)) with iota = -ln(1-omega); omega is never assigned a value and is not ablated, yet it defines the depth supervision target.
  • loss weights gamma, eta, beta, phi = gamma=0.1, eta=1, beta=0.1, phi=0.8
    Fixed across datasets in Section IV-B; treated as tuned hyperparameters rather than derived quantities.
  • TV thresholds tau_edge, tau_smooth = 1e-2, 1e-4
    Set empirically in Section III-D to filter strong edges and small gradients.
  • Canny thresholds = 20, 200
    Lower and upper Canny edge-detection thresholds, empirically set in Section III-C.
  • patch size range = [5, 20]
    Uniformly sampled patch sizes for hierarchical depth supervision; ablation in Table V shows low sensitivity, so this is a minor free parameter.
  • error tolerance mechanism = undefined
    Mentioned in Section IV-B ('we introduce an error tolerance mechanism into the depth L2 loss') but no formula or value is provided.
assumptions (4)
  • domain assumption Monocular depth priors from Depth Anything V2 are reliable geometric guidance for 3DGS training.
    The Ldepth loss pushes rendered depth toward Depth Anything V2 predictions; the paper's limitation section notes these estimates can be wrong in texture-less or occluded regions.
  • domain assumption Canny edges on ground-truth RGB images align with depth discontinuities.
    The edge-aware depth regularization uses Canny masks to gate smoothing; texture edges that do not correspond to geometry would be incorrectly protected from smoothing.
  • domain assumption Ground-truth RGB gradient magnitude indicates where TV smoothing is safe.
    The RGB-guided TV loss assumes small ground-truth gradients identify homogeneous regions suitable for smoothing; lighting or texture variation can violate this.
  • ad hoc to paper The modified depth rendering in Eq. 4 with geometric decay weighting is a valid depth proxy for optimization.
    This weighting is introduced specifically to emphasize near Gaussians; it differs from standard alpha-composited depth and its parameter omega is not specified.

how reviews work

0 comments
Cite this review

Pith. "Pith review of DET-GS: Depth- and Edge-Aware Regularization for High-Fidelity 3D Gaussian Splatting." pith.science (2026). https://pith.science/paper/SDZVQHK7

@misc{pith2026250804099,
  author       = {Pith},
  title        = {Pith review of: DET-GS: Depth- and Edge-Aware Regularization for High-Fidelity 3D Gaussian Splatting},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/SDZVQHK7}},
  note         = {Machine review of arXiv:2508.04099}
}
read the original abstract

3D Gaussian Splatting (3DGS) represents a significant advancement in the field of efficient and high-fidelity novel view synthesis. Despite recent progress, achieving accurate geometric reconstruction under sparse-view conditions remains a fundamental challenge. Existing methods often rely on non-local depth regularization, which fails to capture fine-grained structures and is highly sensitive to depth estimation noise. Furthermore, traditional smoothing methods neglect semantic boundaries and indiscriminately degrade essential edges and textures, consequently limiting the overall quality of reconstruction. In this work, we propose DET-GS, a unified depth and edge-aware regularization framework for 3D Gaussian Splatting. DET-GS introduces a hierarchical geometric depth supervision framework that adaptively enforces multi-level geometric consistency, significantly enhancing structural fidelity and robustness against depth estimation noise. To preserve scene boundaries, we design an edge-aware depth regularization guided by semantic masks derived from Canny edge detection. Furthermore, we introduce an RGB-guided edge-preserving Total Variation loss that selectively smooths homogeneous regions while rigorously retaining high-frequency details and textures. Extensive experiments demonstrate that DET-GS achieves substantial improvements in both geometric accuracy and visual fidelity, outperforming state-of-the-art (SOTA) methods on sparse-view novel view synthesis benchmarks.

Figures

Figures reproduced from arXiv: 2508.04099 by the authors.

Figure 1
Figure 1. DET-GS: Our method reconstructs 3D scenes from multi-view RGB inputs using a representation of 3D Gaussian primitives. To enhance geometric [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Overview of DET-GS: Given a set of RGB images, a pre-trained monocular depth estimator provides monocular depth priors to guide the optimization. The scene is represented by a set of 3D Gaussian primitives, which can render both color images and depth maps. During training, we decompose both the predicted and pseudo-ground-truth depth maps into non-overlapping patches and apply Hierarchical Geometric Depth Supervisi… view at source ↗
Figure 3
Figure 3. Qualitative Comparison Results Across Different Datasets [20], [58]. Zoomed-in regions showcase fine-grained rendering differences. Red arrows mark visual artifacts, such as local blurriness, present in current state-of-the-art [8]–[10] approaches. Compared to these methods, our model achieves more faithful detail preservation and delivers more realistic and high-fidelity renderings. representative methods: Instant-… view at source ↗
Figures from the paper (3 more)
Figure 4
Figure 4. Figure 4: Qualitative Comparison Results of NeRF Synthetic Dataset [1]. The red boxes highlight detailed regions of the rendered images. With the aid of depth and edge-aware supervision, as well as our proposed TV loss, our approach better preserves intricate visual structures a…
Figure 5
Figure 5. Figure 5: Ablation of Our Model Components. We conduct experiments that remove each of the three core innovations individually to assess their effectiveness. We present the PSNR and LPIPS metrics, with the best scores highlighted in red. methods, including point-based renderers …
Figure 6
Figure 6. Figure 6: Depth Map Comparisons Between Our Method and State-of-the-art models [10]. While Spec-Gaussian exhibits blurry floaters and less accurate geometric details, our model generates sharper and structurally consistent depth maps. This improvement is attributed to our Hierar…

Discussion (0). Sign in to comment.

Forward citations

Cited by 2 Pith papers

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

  1. ConFixGS: Learning to Fix Feedforward 3D Gaussian Splatting with Confidence-Aware Diffusion Priors in Driving Scenes

    cs.CV 2026-05 unverdicted novelty 7.0 of 10

    ConFixGS repairs feedforward 3D Gaussian Splatting with confidence-aware diffusion priors, delivering up to 3.68 dB PSNR gains and halved FID scores on Waymo, nuScenes, and KITTI novel view synthesis tasks.

  2. EnerGS: Energy-Based Gaussian Splatting with Partial Geometric Priors

    cs.CV 2026-04 unverdicted novelty 6.0 of 10

    EnerGS introduces an energy-based soft guidance mechanism for partial geometry in 3D Gaussian Splatting to improve reconstruction quality and reduce overfitting in sparse outdoor settings.

Reference graph

Works this paper leans on

63 extracted references · 50 canonical work pages · cited by 2 Pith papers

  1. [1]

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

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

    2021

  2. [2]

    Mvsnerf: Fast generalizable radiance field reconstruction from multi-view stereo,

    A. Chen, Z. Xu, F. Zhao, X. Zhang, F. Xiang, J. Yu, and H. Su, “Mvsnerf: Fast generalizable radiance field reconstruction from multi-view stereo,” in Proceedings of the IEEE/CVF international conference on computer vision, 2021, pp. 14 124–14 133

    work page 2021

  3. [3]

    Freenerf: Improving few-shot neural rendering with free frequency regularization,

    J. Yang, M. Pavone, and Y . Wang, “Freenerf: Improving few-shot neural rendering with free frequency regularization,” in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , 2023, pp. 8254–8263

    work page 2023

  4. [4]

    pixelnerf: Neural radiance fields from one or few images,

    A. Yu, V . Ye, M. Tancik, and A. Kanazawa, “pixelnerf: Neural radiance fields from one or few images,” in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , 2021, pp. 4578– 4587. 11

    work page 2021

  5. [5]

    Instant neural graphics primitives with a multiresolution hash encoding,

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

    work page 2022

  6. [6]

    Direct voxel grid optimization: Super- fast convergence for radiance fields reconstruction,

    C. Sun, M. Sun, and H.-T. Chen, “Direct voxel grid optimization: Super- fast convergence for radiance fields reconstruction,” in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , 2022, pp. 5459–5469

    work page 2022

  7. [7]

    3d gaussian splatting for real-time radiance field rendering

    B. Kerbl, G. Kopanas, T. Leimk ¨uhler, and G. Drettakis, “3d gaussian splatting for real-time radiance field rendering.” ACM Trans. Graph. , vol. 42, no. 4, pp. 139–1, 2023

    2023

  8. [8]

    Mip-splatting: Alias- free 3d gaussian splatting,

    Z. Yu, A. Chen, B. Huang, T. Sattler, and A. Geiger, “Mip-splatting: Alias- free 3d gaussian splatting,” in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , 2024, pp. 19 447–19 456

    work page 2024

Show all 63 references
  1. [9]

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

    T. Lu, M. Yu, L. Xu, Y . Xiangli, L. Wang, D. Lin, and B. Dai, “Scaffold-gs: Structured 3d gaussians for view-adaptive rendering,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2024, pp. 20 654–20 664

    2024

  2. [10]

    Spec-gaussian: Anisotropic view-dependent appearance for 3d gaussian splatting,

    Z. Yang, X. Gao, Y .-T. Sun, Y . Huang, X. Lyu, W. Zhou, S. Jiao, X. Qi, and X. Jin, “Spec-gaussian: Anisotropic view-dependent appearance for 3d gaussian splatting,” Advances in Neural Information Processing Systems, vol. 37, pp. 61 192–61 216, 2024

    2024

  3. [11]

    Depth-regularized optimization for 3d gaussian splatting in few-shot images,

    J. Chung, J. Oh, and K. M. Lee, “Depth-regularized optimization for 3d gaussian splatting in few-shot images,” in 2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW) , 2024, pp. 811–820

    2024

  4. [12]

    Lm-gaussian: Boost sparse-view 3d gaussian splatting with large model priors,

    H. Yu, X. Long, and P. Tan, “Lm-gaussian: Boost sparse-view 3d gaussian splatting with large model priors,” arXiv preprint arXiv:2409.03456 , 2024

    2024 arXiv

  5. [13]

    Dof-gaussian: Controllable depth-of-field for 3d gaussian splatting,

    L. Shen, T. Liu, H. Sun, J. Li, Z. Cao, W. Li, and C. C. Loy, “Dof-gaussian: Controllable depth-of-field for 3d gaussian splatting,” in Proceedings of the Computer Vision and Pattern Recognition Conference , 2025, pp. 26 462–26 471

    2025

  6. [14]

    Dngaussian: Optimizing sparse-view 3d gaussian radiance fields with global-local depth normalization,

    J. Li, J. Zhang, X. Bai, J. Zheng, X. Ning, J. Zhou, and L. Gu, “Dngaussian: Optimizing sparse-view 3d gaussian radiance fields with global-local depth normalization,” in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , 2024, pp. 20 775– 20 785

    2024

  7. [15]

    A computational approach to edge detection,

    J. Canny, “A computational approach to edge detection,” IEEE Transac- tions on pattern analysis and machine intelligence , no. 6, pp. 679–698, 1986

    1986

  8. [16]

    Deepsdf: Learning continuous signed distance functions for shape representation,

    J. J. Park, P. Florence, J. Straub, R. Newcombe, and S. Lovegrove, “Deepsdf: Learning continuous signed distance functions for shape representation,” in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , 2019, pp. 165–174

    2019

  9. [17]

    Neus: Learning neural implicit surfaces by volume rendering for multi-view reconstruction,

    P. Wang, L. Liu, Y . Liu, C. Theobalt, T. Komura, and W. Wang, “Neus: Learning neural implicit surfaces by volume rendering for multi-view reconstruction,” NeurIPS, 2021

    2021

  10. [18]

    Neus2: Fast learning of neural implicit surfaces for multi- view reconstruction,

    Y . Wang, Q. Han, M. Habermann, K. Daniilidis, C. Theobalt, and L. Liu, “Neus2: Fast learning of neural implicit surfaces for multi- view reconstruction,” in Proceedings of the IEEE/CVF International Conference on Computer Vision , 2023, pp. 3295–3306

    2023

  11. [19]

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

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

    2021

  12. [20]

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

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

    2022

  13. [21]

    Plenoctrees for real-time rendering of neural radiance fields,

    A. Yu, R. Li, M. Tancik, H. Li, R. Ng, and A. Kanazawa, “Plenoctrees for real-time rendering of neural radiance fields,” in Proceedings of the IEEE/CVF international conference on computer vision , 2021, pp. 5752–5761

    2021

  14. [22]

    Neural sparse voxel fields,

    L. Liu, J. Gu, K. Zaw Lin, T.-S. Chua, and C. Theobalt, “Neural sparse voxel fields,” Advances in Neural Information Processing Systems , vol. 33, pp. 15 651–15 663, 2020

    2020

  15. [23]

    Plenoxels: Radiance fields without neural networks,

    S. Fridovich-Keil, A. Yu, M. Tancik, Q. Chen, B. Recht, and A. Kanazawa, “Plenoxels: Radiance fields without neural networks,” in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , 2022, pp. 5501–5510

    2022

  16. [24]

    Tensorf: Tensorial radiance fields,

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

    2022

  17. [25]

    Efficient neural implicit representation for 3d human reconstruction,

    Z. Huang, S. M. Erfani, S. Lu, and M. Gong, “Efficient neural implicit representation for 3d human reconstruction,” Pattern Recognition, vol. 156, p. 110758, 2024

    2024

  18. [26]

    Vr-nerf: High-fidelity virtualized walkable spaces,

    L. Xu, V . Agrawal, W. Laney, T. Garcia, A. Bansal, C. Kim, S. Rota Bul`o, L. Porzi, P. Kontschieder, A. Bo ˇziˇc et al. , “Vr-nerf: High-fidelity virtualized walkable spaces,” in SIGGRAPH Asia 2023 Conference Papers, 2023, pp. 1–12

    2023

  19. [27]

    Zip-nerf: Anti-aliased grid-based neural radiance fields,

    J. T. Barron, B. Mildenhall, D. Verbin, P. P. Srinivasan, and P. Hedman, “Zip-nerf: Anti-aliased grid-based neural radiance fields,” in Proceedings of the IEEE/CVF International Conference on Computer Vision , 2023, pp. 19 697–19 705

    2023

  20. [28]

    Kitti-360: A novel dataset and benchmarks for urban scene understanding in 2d and 3d,

    Y . Liao, J. Xie, and A. Geiger, “Kitti-360: A novel dataset and benchmarks for urban scene understanding in 2d and 3d,” IEEE Transactions on Pattern Analysis and Machine Intelligence , vol. 45, no. 3, pp. 3292–3310, 2022

    2022

  21. [29]

    Pixelwise view selection for unstructured multi-view stereo,

    J. L. Sch ¨onberger, E. Zheng, J.-M. Frahm, and M. Pollefeys, “Pixelwise view selection for unstructured multi-view stereo,” in Computer Vision– ECCV 2016: 14th European Conference, Amsterdam, The Netherlands, October 11-14, 2016, Proceedings, Part III 14 . Springer, 2016, pp. 501–518

    2016

  22. [30]

    Adop: Approximate differentiable one-pixel point rendering,

    D. R ¨uckert, L. Franke, and M. Stamminger, “Adop: Approximate differentiable one-pixel point rendering,” ACM Transactions on Graphics (ToG), vol. 41, no. 4, pp. 1–14, 2022

    2022

  23. [31]

    Point-based neural rendering with per-view optimization,

    G. Kopanas, J. Philip, T. Leimk ¨uhler, and G. Drettakis, “Point-based neural rendering with per-view optimization,” in Computer Graphics F orum, vol. 40, no. 4. Wiley Online Library, 2021, pp. 29–43

    2021

  24. [32]

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

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

    2020

  25. [33]

    Broad-to- narrow registration and identification of 3d objects in partially scanned and cluttered point clouds,

    G. Arvanitis, E. I. Zacharaki, L. V ´aˆsa, and K. Moustakas, “Broad-to- narrow registration and identification of 3d objects in partially scanned and cluttered point clouds,” IEEE Transactions on Multimedia , vol. 24, pp. 2230–2245, 2021

    2021

  26. [34]

    Structgs: Adaptive spherical harmonics and rendering enhancements for superior 3d gaussian splatting,

    Z. Huang, M. Xu, and S. Perry, “Structgs: Adaptive spherical harmonics and rendering enhancements for superior 3d gaussian splatting,” arXiv preprint arXiv:2503.06462, 2025

    2025

  27. [35]

    Gaussianfocus: Constrained attention focus for 3d gaussian splatting,

    ——, “Gaussianfocus: Constrained attention focus for 3d gaussian splatting,” arXiv preprint arXiv:2503.17798 , 2025

    2025

  28. [36]

    Motion-aware 3d gaussian splatting for efficient dynamic scene reconstruction,

    Z. Guo, W. Zhou, L. Li, M. Wang, and H. Li, “Motion-aware 3d gaussian splatting for efficient dynamic scene reconstruction,” IEEE Transactions on Circuits and Systems for Video Technology , 2024

    2024

  29. [37]

    Gedr: Gaussian- enhanced detail reconstruction for real-time high-fidelity 3d scene reconstruction,

    T. Zhou, S. Chen, S. Wan, H. Lv, Z. Luo, and J. Wu, “Gedr: Gaussian- enhanced detail reconstruction for real-time high-fidelity 3d scene reconstruction,” IEEE Transactions on Circuits and Systems for Video Technology, 2025

    2025

  30. [38]

    Frpgs: Fast, robust, and photorealistic monocular dynamic scene reconstruction with deformable 3d gaussians,

    W. Li, X. Pan, J. Lin, P. Lu, D. Feng, and W. Shi, “Frpgs: Fast, robust, and photorealistic monocular dynamic scene reconstruction with deformable 3d gaussians,” IEEE Transactions on Circuits and Systems for Video Technology, 2025

    2025

  31. [39]

    Get3dgs: Generate 3d gaussians based on points deformation fields,

    H. Yu, W. Gong, J. Chen, and H. Ma, “Get3dgs: Generate 3d gaussians based on points deformation fields,” IEEE Transactions on Circuits and Systems for Video Technology , 2024

    2024

  32. [40]

    Gps-gaussian: Generalizable pixel-wise 3d gaussian splatting for real- time human novel view synthesis,

    S. Zheng, B. Zhou, R. Shao, B. Liu, S. Zhang, L. Nie, and Y . Liu, “Gps-gaussian: Generalizable pixel-wise 3d gaussian splatting for real- time human novel view synthesis,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) , 2024

    2024

  33. [41]

    3dgs-avatar: Animatable avatars via deformable 3d gaussian splatting,

    Z. Qian, S. Wang, M. Mihajlovic, A. Geiger, and S. Tang, “3dgs-avatar: Animatable avatars via deformable 3d gaussian splatting,” in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , 2024, pp. 5020–5030

    2024

  34. [42]

    Multi-scale 3d gaussian splatting for anti-aliased rendering,

    Z. Yan, W. F. Low, Y . Chen, and G. H. Lee, “Multi-scale 3d gaussian splatting for anti-aliased rendering,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , 2024, pp. 20 923–20 931

    2024

  35. [43]

    Mesh-aligned 3d gaussian splatting for multi-resolution anti-aliasing rendering,

    J. Liu, L. Kong, J. Yan, and G. Chen, “Mesh-aligned 3d gaussian splatting for multi-resolution anti-aliasing rendering,” IEEE Transactions on Circuits and Systems for Video Technology , 2025

    2025

  36. [44]

    Uncertainty estimation for stereo matching based on evidential deep learning,

    C. Wang, X. Wang, J. Zhang, L. Zhang, X. Bai, X. Ning, J. Zhou, and E. Hancock, “Uncertainty estimation for stereo matching based on evidential deep learning,” pattern recognition, vol. 124, p. 108498, 2022

    2022

  37. [45]

    Robust training for multi-view stereo networks with noisy labels,

    X. Wang, H. Luo, Z. Wang, J. Zheng, and X. Bai, “Robust training for multi-view stereo networks with noisy labels,” Displays, vol. 81, p. 102604, 2024

    2024

  38. [46]

    A contrastive learning based unsupervised multi-view stereo with multi-stage self- training strategy,

    Z. Wang, H. Luo, X. Wang, J. Zheng, X. Ning, and X. Bai, “A contrastive learning based unsupervised multi-view stereo with multi-stage self- training strategy,” Displays, vol. 83, p. 102672, 2024

    2024

  39. [47]

    Depth-supervised nerf: Fewer views and faster training for free,

    K. Deng, A. Liu, J.-Y . Zhu, and D. Ramanan, “Depth-supervised nerf: Fewer views and faster training for free,” in Proceedings of the IEEE/CVF 12 conference on computer vision and pattern recognition , 2022, pp. 12 882– 12 891

    2022

  40. [48]

    Dense depth priors for neural radiance fields from sparse input views,

    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 Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2022, pp. 12 892–12 901

    2022

  41. [49]

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

    J. Song, S. Park, H. An, S. Cho, M.-S. Kwak, S. Cho, and S. Kim, “D¨arf: Boosting radiance fields from sparse input views with monocular depth adaptation,” Advances in Neural Information Processing Systems , vol. 36, pp. 68 458–68 470, 2023

    2023

  42. [50]

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

    G. Wang, Z. Chen, C. C. Loy, and Z. Liu, “Sparsenerf: Distilling depth ranking for few-shot novel view synthesis,” in Proceedings of the IEEE/CVF international conference on computer vision , 2023, pp. 9065–9076

    2023

  43. [51]

    Monosdf: Exploring monocular geometric cues for neural implicit surface recon- struction,

    Z. Yu, S. Peng, M. Niemeyer, T. Sattler, and A. Geiger, “Monosdf: Exploring monocular geometric cues for neural implicit surface recon- struction,” Advances in neural information processing systems , vol. 35, pp. 25 018–25 032, 2022

    2022

  44. [52]

    Vision transformers for dense prediction,

    R. Ranftl, A. Bochkovskiy, and V . Koltun, “Vision transformers for dense prediction,” in Proceedings of the IEEE/CVF international conference on computer vision , 2021, pp. 12 179–12 188

    2021

  45. [53]

    Depth anything: Unleashing the power of large-scale unlabeled data,

    L. Yang, B. Kang, Z. Huang, X. Xu, J. Feng, and H. Zhao, “Depth anything: Unleashing the power of large-scale unlabeled data,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2024, pp. 10 371–10 381

    2024

  46. [54]

    Depth anything v2,

    L. Yang, B. Kang, Z. Huang, Z. Zhao, X. Xu, J. Feng, and H. Zhao, “Depth anything v2,” Advances in Neural Information Processing Systems , vol. 37, pp. 21 875–21 911, 2024

    2024

  47. [55]

    Xiong, SparseGS: Real-time 360 ° sparse view synthesis using Gaussian splatting

    H. Xiong, SparseGS: Real-time 360 ° sparse view synthesis using Gaussian splatting . University of California, Los Angeles, 2024

    2024

  48. [56]

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

    Z. Zhu, Z. Fan, Y . Jiang, and Z. Wang, “Fsgs: Real-time few-shot view synthesis using gaussian splatting,” in European conference on computer vision. Springer, 2024, pp. 145–163

    2024

  49. [57]

    Structure-from-motion revisited,

    J. L. Schonberger and J.-M. Frahm, “Structure-from-motion revisited,” in Proceedings of the IEEE conference on computer vision and pattern recognition, 2016, pp. 4104–4113

    2016

  50. [58]

    Tanks and temples: Benchmarking large-scale scene reconstruction,

    A. Knapitsch, J. Park, Q.-Y . Zhou, and V . Koltun, “Tanks and temples: Benchmarking large-scale scene reconstruction,” ACM Transactions on Graphics (ToG), vol. 36, no. 4, pp. 1–13, 2017

    2017

  51. [59]

    Deep blending for free-viewpoint image-based rendering,

    P. Hedman, J. Philip, T. Price, J.-M. Frahm, G. Drettakis, and G. Bros- tow, “Deep blending for free-viewpoint image-based rendering,” ACM Transactions on Graphics (ToG) , vol. 37, no. 6, pp. 1–15, 2018

    2018

  52. [60]

    Tri-miprf: Tri-mip representation for efficient anti-aliasing neural radiance fields,

    W. Hu, Y . Wang, L. Ma, B. Yang, L. Gao, X. Liu, and Y . Ma, “Tri-miprf: Tri-mip representation for efficient anti-aliasing neural radiance fields,” in Proceedings of the IEEE/CVF International Conference on Computer Vision, 2023, pp. 19 774–19 783

    2023

  53. [61]

    Neurbf: A neural fields representation with adaptive radial basis functions,

    Z. Chen, Z. Li, L. Song, L. Chen, J. Yu, J. Yuan, and Y . Xu, “Neurbf: A neural fields representation with adaptive radial basis functions,” in Proceedings of the IEEE/CVF International Conference on Computer Vision, 2023, pp. 4182–4194

    2023

  54. [62]

    Image quality assessment: from error visibility to structural similarity,

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

    2004

  55. [63]

    The unreasonable effectiveness of deep features as a perceptual metric,

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

    2018

Pith tools

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