Pith. sign in

REVIEW 4 major objections 5 minor 38 references

K-Buffers: A Plug-in Method for Enhancing Neural Fields with Multiple Buffers

T0 review · 4 major / 5 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read This paper claims that rasterizing K ordered z-buffers and fusing their feature maps with a tiny K-Feature Fusion Network improves rendering quality for neural point fields and 3D Gaussian Splatting while reducing storage.

desk verdict A plausible depth-peeling extension for point-based neural rendering that earns a serious referee, but the advertised 3DGS storage reduction is not backed by the reported per-Gaussian arithmetic. read the letter →

arxiv 2505.19564 v1 pith:RCYC536I submitted 2025-05-26 cs.CV

Haofan Ren , Zunjie Zhu , Xiang Chen , Ming Lu , Rongfeng Lu , Chenggang Yan This is my paper · ORCID
classification cs.CV
keywords neuralfields3DGaussianSplattingpointz-bufferK-FeatureFusionNetworknovelviewsynthesiscloudrenderingrasterization
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 sets out to show that the rendering process, not just the scene representation, can be upgraded: instead of shading each pixel from the single nearest surface point, it rasterizes the K nearest depth layers, turns each layer into a pixel-wise feature map, and merges the K maps in latent space. This targets the known weakness of point-based renderers: noisy point clouds and hole-ridden rasterized images. If the claim holds, K-Buffers works as a plug-in that lifts PSNR and SSIM and lowers LPIPS for neural point fields and 3D Gaussian Splatting, and it can cut 3DGS storage substantially. The authors report consistent gains on NeRF-Synthetic, ScanNet, DTU, and Mip-NeRF360, with the largest storage drop on NeRF-Synthetic from 170.5 to 34.87 MB.

What carries the argument

The machinery is the K z-buffer itself plus three small components: a pruning rule that keeps only the direction $d_j$ with the smallest pixel ID $j = \min \mathcal{A}_{p_i}$ for each 3D point, so storing K layers does not multiply the number of neural points; a feature rectification term $T_\Psi(o, d_j)$ added to the radiance mapping $F_\Theta(x_m, d_m)$; and KFN, a two-convolution-layer network with softmax normalization that predicts pixel-wise scalar masks in $[0,1]$ to merge the K feature maps before a U-Net decodes the fused features to color.

What would settle it

Re-run the 3DGS variant with K=1 while keeping KFN fusion: if PSNR does not drop relative to K=8, the improvement comes from fusion, not multiple depth layers. Separately, keep the original spherical-harmonic colors and add only the K-Buffer fusion: if the storage figures do not fall from 170.5 MB toward 34.87 MB on NeRF-Synthetic, the storage reduction is a representation swap, not the K-Buffers mechanism.

Watch

Extended reading notes

Core claim

The central claim is that keeping K ordered z-buffers during the depth test, encoding each surviving queried point through a radiance-mapping feature extractor, pruning redundant point directions so that one 3D point generates at most one query, and fusing the resulting K feature maps with a tiny K-Feature Fusion Network yields cleaner, higher-fidelity renders than the standard single-buffer rasterization pipeline. A naive alternative, volume-integrating the K predicted colors, overfits, so the discovery is that fusion must happen in latent feature space, where KFN's softmax-normalized masks can denoise across layers. Applied as a plug-in, the same recipe improves the rendering quality of neural point fields and 3D Gaussian Splatting, and it reduces the storage footprint of 3DGS in the reported experiments.

Load-bearing premise

The load-bearing assumption is that the 3DGS comparison is apples-to-apples; the paper replaces the usual per-Gaussian color encoding with a 32-dim feature plus a small MLP, and if that swap, rather than the K-Buffer fusion, is what shrinks storage from 170.5 MB to 34.87 MB, the storage claim does not test what it appears to test.

Editorial extensions

If this is right

  • Neural point field baselines BPCR and FrePCR both gain rendering quality when K-Buffers is added; for example BPCR PSNR rises from 29.12 to 29.58 on NeRF-Synthetic and from 25.88 to 26.66 on ScanNet.
  • 3DGS rendering quality improves as well, with PSNR on NeRF-Synthetic going from 33.37 to 33.62 and on Mip-NeRF360 from 28.93 to 29.19, while LPIPS on Mip-NeRF360 drops from 0.136 to 0.126.
  • 3DGS storage drops substantially in the reported experiments, from 735.57 MB to 383.43 MB on Mip-NeRF360 and from 170.50 MB to 34.87 MB on NeRF-Synthetic.
  • Relative to the unpruned K=8 variant (1,613,013 queried points, 386.86 GFLOPS, 33 GB), the pruning strategy cuts the count to 194,225, GFLOPS to 122.43, and memory to 7 GB, a more than threefold reduction.
  • The loss function is unchanged from the baseline, so K-Buffers modifies only the rendering side, and temporal consistency of 3DGS is reported as not compromised under MSE and FLIP evaluation.

Reading between the lines

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

  • A natural extension the paper does not test is applying the same K-Buffer fusion to other rasterization renderers, such as ADOP or NPBG++; if the mechanism is generic, similar gains should appear whenever point splatting leaves holes or noise.
  • The storage numbers for 3DGS mix two changes, the K-Buffer pipeline and the replacement of the spherical-harmonic color encoding by a 32-dim feature plus a small MLP, so an ablation that holds the color representation fixed would isolate how much storage K-Buffers itself saves.
  • Because the paper reports quality rising with K while FPS falls, a view-adaptive or scene-adaptive choice of K could push the speed-quality frontier beyond the fixed K=8 setting used in the experiments.
  • The success of softmax-mask fusion in latent space over K depth layers suggests a direct connection to burst denoising; one could test whether a per-pixel weighted average in color space, rather than feature space, loses the reported denoising benefit.
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 K-Buffers, a plug-in rendering enhancement for point-based neural fields and 3D Gaussian Splatting. The method renders K z-buffer layers instead of a single depth map, constructs K pixel-wise feature maps, prunes redundant queried points, rectifies features with a small MLP, and fuses the K feature maps with a K-Feature Fusion Network before decoding with a U-Net. Experiments on NeRF-Synthetic, ScanNet, DTU, and Mip-NeRF 360 report PSNR/SSIM/LPIPS gains for multiple baselines and claim substantial 3DGS storage reductions. The paper also includes ablations of the pruning and fusion components and a temporal-consistency evaluation.

Significance. If the rendering-quality improvements are reproducible, the K-Buffers idea is a useful and simple plug-in for rasterization-based neural rendering: it directly targets the noise and hole artifacts caused by single-layer depth tests, and the proposed pruning strategy is an interesting way to control the cost of multiple buffers. The authors provide code, ablate the main components, and evaluate across several benchmarks. However, the advertised 3DGS storage reduction is not supported by the reported representation change, the 3DGS variant of the method is underspecified, and the DTU protocol is altered in a way that makes baseline comparisons unreliable. These issues need to be addressed before the central claims can be accepted.

major comments (4)
  1. [Abstract; §3.3; Tables 20–22] The storage reduction advertised for 3DGS is not supported by the reported representation change. In §3.3 (Eqs. 6–7) you replace the 48 SH coefficients of standard 3DGS with a 32-dimensional feature, reducing per-Gaussian storage from about 59 to 43 floats (a factor of 1.37), yet Tables 20–22 report reductions of about 4.9x on NeRF-Synthetic (170.50 to 34.87 MB) and similar on ScanNet/DTU. The missing Gaussian counts and per-Gaussian parameter layouts mean the reported drop could come from a different number of Gaussians or from other unstated changes, not from the K-Buffers/KFN mechanism. Please report Gaussian counts, per-Gaussian byte layout, and the size of the added MLP/KFN/U-Net, and either include those network weights in 'Storage' or explicitly state that they are excluded.
  2. [§3.3, Eq. (6), Table 7] The 3DGS variant is underspecified: it is not explained how K z-buffers are obtained from a Gaussian representation, how Eq. (6) (which sums over all N ordered Gaussians) relates to K, or how KFN receives K feature maps in this case. Moreover, Table 7 shows that on NeRF-Synthetic 3DGS(KFN) alone is worse than 3DGS (33.20 vs 33.37) and on DTU it is also worse (33.66 vs 33.91); the final gain appears only after adding the feature rectification/MLP representation change. This weakens the claim that the observed 3DGS improvements are due to the K-Buffers plug-in rather than the altered per-Gaussian representation. Please clarify the K-buffer construction for 3DGS and provide an ablation that isolates the K-Buffers mechanism from the feature/MLP change.
  3. [§A.3, Tables 2 and 14] The DTU evaluation protocol is changed by masking out the background with IDR masks during both training and testing, and the paper states that 'all methods undergo evaluation under these settings.' No evidence is given that the baseline numbers (BPCR, FrePCR, 3DGS) were re-run under the masked protocol; if they were taken from the original papers, the FrePCR baseline of 24.61 PSNR compared with 30.61 after adding your method is not a controlled comparison. Please report re-evaluated baselines under the identical masked protocol, or keep the original protocol and discuss the effect of the change.
  4. [Tables 1, 2, and 7] The claim that K-Buffers 'enhances rendering performance' is not uniformly supported: on ScanNet, BPCR+Ours SSIM drops from 0.794 to 0.789 and 3DGS+Ours LPIPS worsens from 0.403 to 0.406; on DTU, BPCR+Ours LPIPS worsens from 0.128 to 0.151 and 3DGS+Ours LPIPS from 0.050 to 0.059; on Mip-NeRF 360, 3DGS+Ours SSIM drops from 0.869 to 0.859 (Table 1). No error bars or repeated runs are reported, so it is unclear which differences are significant. Please report variance estimates or per-seed results and discuss the regressions explicitly.
minor comments (5)
  1. [§3.1 and §C] Equation numbers are inconsistent: Eq. (4) and Eq. (9) are identical, and Eq. (9) is referenced in §3.1 before it is defined; please renumber.
  2. [§3.2 and §3.3] There are typos: 'direly' should be 'directly' in §3.2, and 'potision' should be 'position' in §3.3.
  3. [Table 6] The caption of Table 6 says FPS increases with K, but the reported FPS decreases from 51.00 (K=1) to 27.52 (K=8); the arrow in the header should be corrected and the text should match the table.
  4. [Figures 7 and 8] Figures 7 and 8 have identical captions ('We assigned colors...'), and Fig. 7's caption refers to z-buffers while Fig. 8 appears to show ScanNet results; the captions should be differentiated.
  5. [Table 3] The note 'The storage includes the point cloud size' is ambiguous for the 3DGS rows, since standard 3DGS does not retain a separate point cloud; please clarify what is included in the reported storage for each method.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: all claims are empirical comparisons against external baselines with ablations; the storage-reduction attribution is a measurement-design concern, not a reduction-by-construction.

full rationale

The paper's central claims are empirical: rendering K z-buffers and fusing the resulting K pixel-wise feature maps with KFN improves PSNR/SSIM/LPIPS on NeRF-Synthetic, ScanNet, DTU, and Mip-NeRF 360 relative to BPCR, FrePCR, and 3DGS baselines. No equation defines the reported metric in terms of the method's fitted parameters; the loss is stated to remain exactly consistent with the baseline, and each added component (KFN, pruning, feature rectification, decoder) is ablated on the same external benchmarks. The reported 3DGS storage reduction (e.g., 170.50 to 34.87 MB in Table 20) is a measured property of the modified 3DGS representation in Eqs. 6-7 rather than a quantity forced by construction, so whether the decrease is attributable to K-Buffers or to the feature-plus-MLP replacement of SH coefficients is a comparison-design question, not circularity. The cited works are external baselines and technique sources (BPCR, FrePCR, 3DGS, KPN, gsplat), and no load-bearing self-citation or imported uniqueness theorem is used. There is therefore no self-definitional step, no fitted input renamed as a prediction, and no ansatz smuggled in via citation.

Assumptions & free parameters 5 free parameters · 6 assumptions · 0 invented entities

The method rests on standard rasterization assumptions plus several design choices (K, C, tau, pruning direction, DTU masking). No new physical entities are postulated. The main unproven premises are that multi-layer z-buffers carry complementary information and that the pruning step preserves needed signal.

free parameters (5)
  • K (number of z-buffer layers) = 8
    Chosen via ablation (Table 6); trades quality versus speed and model size, not derived from theory.
  • Feature dimension C = 8 for neural point fields, 32 for 3DGS
    Hand-chosen; controls capacity of the pixel-wise feature maps.
  • Point radius tau = 5e-3 for NeRF-Synthetic, 1.5e-2 for ScanNet, 3e-3 for DTU
    Set per dataset in Appendix B.2; controls the splat size during rasterization.
  • Positional encoding frequency L = 10 for position, 4 for direction
    Chosen following NeRF; affects the expressiveness of the radiance mapping.
  • Pruning direction selection = dm = direction of minimum pixel ID
    Design choice for the acceleration strategy; tested against random and average choices in Table 9.
assumptions (6)
  • ad hoc to paper A single 3D point expanded to a disk can be represented by one queried point (minimum pixel ID direction) without loss of needed information
    Eq. 4 and Eq. 9; the pruning strategy assumes all other directions are redundant for the final rendered image.
  • domain assumption K z-buffers captured during depth test provide complementary multi-layer geometry information
    Central motivation in Section 1 and Figure 2; not proven, only empirically supported.
  • domain assumption The radiance mapping L = F_Theta(x, d) from BPCR is a sufficient feature encoder for point-based rendering
    Eq. 2; inherited from the baseline and used as the feature extractor for neural point fields.
  • domain assumption The loss function identical to the baseline is sufficient to train the added modules
    Section 3.3 states the loss is completely consistent with the baseline; no auxiliary losses are introduced.
  • domain assumption DTU background masking produces fair, comparable results
    Appendix A.3 changes the NPBG++ protocol by masking the background; the validity of this change is assumed.
  • domain assumption Volume rendering's multi-point integration insight transfers to fusing K rasterized feature maps
    Section 1 motivates K-Buffers by analogy to volume rendering, but the fusion is done in latent space by a learned network.

how reviews work

0 comments
Cite this review

Pith. "Pith review of K-Buffers: A Plug-in Method for Enhancing Neural Fields with Multiple Buffers." pith.science (2026). https://pith.science/paper/RCYC536I

@misc{pith2026250519564,
  author       = {Pith},
  title        = {Pith review of: K-Buffers: A Plug-in Method for Enhancing Neural Fields with Multiple Buffers},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/RCYC536I}},
  note         = {Machine review of arXiv:2505.19564}
}
read the original abstract

Neural fields are now the central focus of research in 3D vision and computer graphics. Existing methods mainly focus on various scene representations, such as neural points and 3D Gaussians. However, few works have studied the rendering process to enhance the neural fields. In this work, we propose a plug-in method named K-Buffers that leverages multiple buffers to improve the rendering performance. Our method first renders K buffers from scene representations and constructs K pixel-wise feature maps. Then, We introduce a K-Feature Fusion Network (KFN) to merge the K pixel-wise feature maps. Finally, we adopt a feature decoder to generate the rendering image. We also introduce an acceleration strategy to improve rendering speed and quality. We apply our method to well-known radiance field baselines, including neural point fields and 3D Gaussian Splatting (3DGS). Extensive experiments demonstrate that our method effectively enhances the rendering performance of neural point fields and 3DGS.

Figures

Figures reproduced from arXiv: 2505.19564 by the authors.

Figure 1
Figure 1. As shown in the figure, our method can simultaneously [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. The motivation of our method. (a) Previous neural point fields first render the pixel-wise z-buffers and then use a decoder to generate the image from the pixel-wise feature map. However, they are sensitive to the noisy point cloud. (b) A naive solution is to render K z-buffers and use volume rendering to integrate the K colors. However, this solution achieves overfitted results. (c) Our method uses KFN to integrate… view at source ↗
Figure 3
Figure 3. The Overview pipeline with our model. (a) depicts how to obtain the neural descriptors from noisy point clouds. (b) describe how [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figures from the paper (6 more)
Figure 4
Figure 4. Figure 4: Illustration of the z-buffer defect. The color of the point [PITH_FULL_IMAGE:figures/full_fig_p005_4.png]
Figure 5
Figure 5. Figure 5: We compare the rendered novel views between ours and previous methods. The scenes are, from the top down: [PITH_FULL_IMAGE:figures/full_fig_p006_5.png]
Figure 6
Figure 6. Figure 6: Our method struggles to produce accurate results in re [PITH_FULL_IMAGE:figures/full_fig_p007_6.png]
Figure 7
Figure 7. Figure 7: We assign colors to the z-buffers of different layers based on the point cloud to observe their geometric differences. [PITH_FULL_IMAGE:figures/full_fig_p010_7.png]
Figure 8
Figure 8. Figure 8: We assigned colors to the z-buffers of different layers based on the point cloud to observe their geometric differences. [PITH_FULL_IMAGE:figures/full_fig_p011_8.png]
Figure 9
Figure 9. Figure 9: Radiance mapping changes over training time on Mip-NeRF dataset. [PITH_FULL_IMAGE:figures/full_fig_p012_9.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

38 extracted references · 36 canonical work pages

  1. [1]

    Neural point-based graphics

    [Aliev et al., 2020] Kara-Ali Aliev, Artem Sevastopolsky, Maria Kolos, Dmitry Ulyanov, and Victor Lempitsky. Neural point-based graphics. In Computer Vision–ECCV 2020: 16th European Conference, Glasgow, UK, August 23–28, 2020, Proceedings, Part XXII 16 , pages 696–712. Springer,

    work page 2020

  2. [3]

    Barron, Ben Mildenhall, Dor Verbin, Pratul P

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

    work page 2022

  3. [7]

    Plenoxels: Radiance fields without neural networks

    [Fridovich-Keil et al., 2022] Sara Fridovich-Keil, Alex Yu, Matthew Tancik, Qinhong Chen, Benjamin Recht, and Angjoo Kanazawa. Plenoxels: Radiance fields without neural networks. In CVPR,

    work page 2022

  4. [8]

    Deep burst denoising

    [Godard et al., 2018] Cl´ement Godard, Kevin Matzen, and Matt Uyttendaele. Deep burst denoising. In Proceedings of the European conference on computer vision (ECCV) , pages 538–554,

    work page 2018

  5. [10]

    Deep blending for free-viewpoint image- based rendering

    [Hedman et al., 2018] Peter Hedman, Julien Philip, True Price, Jan-Michael Frahm, George Drettakis, and Gabriel Brostow. Deep blending for free-viewpoint image- based rendering. ACM Transactions on Graphics (ToG) , 37(6):1–15,

    work page 2018

  6. [11]

    Trivol: Point cloud rendering via triple vol- umes

    [Hu et al., 2023] Tao Hu, Xiaogang Xu, Ruihang Chu, and Jiaya Jia. Trivol: Point cloud rendering via triple vol- umes. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages 20732– 20741,

    work page 2023

  7. [12]

    Boosting point clouds rendering via radiance mapping

    [Huang et al., 2023] Xiaoyang Huang, Yi Zhang, Bingbing Ni, Teng Li, Kai Chen, and Wenjun Zhang. Boosting point clouds rendering via radiance mapping. In Proceedings of the AAAI conference on artificial intelligence, volume 37, pages 953–961,

    work page 2023

  8. [13]

    Large scale multi-view stereopsis evaluation

    [Jensen et al., 2014] Rasmus Jensen, Anders Dahl, George V ogiatzis, Engin Tola, and Henrik Aanæs. Large scale multi-view stereopsis evaluation. In Proceedings of the IEEE conference on computer vision and pattern recogni- tion, pages 406–413,

    work page 2014

Show all 38 references
  1. [15]

    Read: Large-scale neural scene rendering for autonomous driv- ing

    [Li et al., 2023] Zhuopeng Li, Lu Li, and Jianke Zhu. Read: Large-scale neural scene rendering for autonomous driv- ing. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 37, pages 1522–1529,

    2023

  2. [16]

    Hashpoint: Accelerated point searching and sampling for neural rendering

    [Ma et al., 2024] Jiahao Ma, Miaomiao Liu, David Ahmedt- Aristizabal, and Chuong Nguyen. Hashpoint: Accelerated point searching and sampling for neural rendering. InPro- ceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 4462–4472,

    2024

  3. [18]

    Srini- vasan, Matthew Tancik, Jonathan T

    [Mildenhall et al., 2020] Ben Mildenhall, Pratul P. Srini- vasan, Matthew Tancik, Jonathan T. Barron, Ravi Ra- mamoorthi, and Ren Ng. Nerf: Representing scenes as neural radiance fields for view synthesis. In ECCV,

    2020

  4. [19]

    Instant neural graphics primitives with a multiresolution hash encoding

    [M¨uller et al., 2022] Thomas M¨uller, Alex Evans, Christoph Schied, and Alexander Keller. Instant neural graphics primitives with a multiresolution hash encoding. ACM Transactions on Graphics (ToG), 41(4):1–15,

    2022

  5. [20]

    Neural point light fields

    [Ost et al., 2022] Julian Ost, Issam Laradji, Alejandro Newell, Yuval Bahat, and Felix Heide. Neural point light fields. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages 18419– 18429,

    2022

  6. [21]

    StopThePop: Sorted Gaussian Splat- ting for View-Consistent Real-time Rendering

    [Radl et al., 2024] Lukas Radl, Michael Steiner, Mathias Parger, Alexander Weinrauch, Bernhard Kerbl, and Markus Steinberger. StopThePop: Sorted Gaussian Splat- ting for View-Consistent Real-time Rendering. ACM Transactions on Graphics, 43(4),

    2024

  7. [22]

    Npbg++: Accelerating neural point-based graphics

    [Rakhimov et al., 2022] Ruslan Rakhimov, Andrei-Timotei Ardelean, Victor Lempitsky, and Evgeny Burnaev. Npbg++: Accelerating neural point-based graphics. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages 15969–15979,

    2022

  8. [23]

    Accelerating 3d deep learning with pytorch3d

    [Ravi et al., 2020] Nikhila Ravi, Jeremy Reizenstein, David Novotny, Taylor Gordon, Wan-Yen Lo, Justin Johnson, and Georgia Gkioxari. Accelerating 3d deep learning with pytorch3d. arXiv:2007.08501,

    2020 arXiv

  9. [24]

    Kilonerf: Speeding up neural radiance fields with thousands of tiny mlps

    [Reiser et al., 2021] Christian Reiser, Songyou Peng, Yiyi Liao, and Andreas Geiger. Kilonerf: Speeding up neural radiance fields with thousands of tiny mlps. In Proceed- ings of the IEEE/CVF International Conference on Com- puter Vision, pages 14335–14345,

    2021

  10. [27]

    Structure-from-motion revisited

    [Sch¨onberger and Frahm, 2016] Johannes Lutz Sch ¨onberger and Jan-Michael Frahm. Structure-from-motion revisited. In Conference on Computer Vision and Pattern Recogni- tion (CVPR),

    2016

  11. [30]

    Patchmatchnet: Learned multi-view patchmatch stereo,

    [Wang et al., 2021] Fangjinhua Wang, Silvano Galliani, Christoph V ogel, Pablo Speciale, and Marc Pollefeys. Patchmatchnet: Learned multi-view patchmatch stereo,

    2021

  12. [31]

    Ba- sis prediction networks for effective burst denoising with large kernels

    [Xia et al., 2020] Zhihao Xia, Federico Perazzi, Micha ¨el Gharbi, Kalyan Sunkavalli, and Ayan Chakrabarti. Ba- sis prediction networks for effective burst denoising with large kernels. In Proceedings of the IEEE/CVF Confer- ence on Computer Vision and Pattern Recognition , pa...

    2020

  13. [32]

    Point-nerf: Point-based neural radiance fields

    [Xu et al., 2022] Qiangeng Xu, Zexiang Xu, Julien Philip, Sai Bi, Zhixin Shu, Kalyan Sunkavalli, and Ulrich Neu- mann. Point-nerf: Point-based neural radiance fields. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 5438–5448,

    2022

  14. [33]

    Multiview neural surface reconstruction by dis- entangling geometry and appearance

    [Yarivet al., 2020] Lior Yariv, Yoni Kasten, Dror Moran, Meirav Galun, Matan Atzmon, Basri Ronen, and Yaron Lipman. Multiview neural surface reconstruction by dis- entangling geometry and appearance. Advances in Neural Information Processing Systems, 33:2492–2502,

    2020

  15. [34]

    Mathematical supplement for the gsplat library,

    [Ye and Kanazawa, 2023] Vickie Ye and Angjoo Kanazawa. Mathematical supplement for the gsplat library,

    2023

  16. [35]

    Plenoctrees for real- time rendering of neural radiance fields

    [Yu et al., 2021] Alex Yu, Ruilong Li, Matthew Tancik, Hao Li, Ren Ng, and Angjoo Kanazawa. Plenoctrees for real- time rendering of neural radiance fields. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 5752–5761,

    2021

  17. [36]

    Mip-splatting: Alias-free 3d gaussian splatting

    [Yu et al., 2024] Zehao Yu, Anpei Chen, Binbin Huang, Torsten Sattler, and Andreas Geiger. Mip-splatting: Alias-free 3d gaussian splatting. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 19447–19456, June

    2024

  18. [37]

    Frequency-modulated point cloud rendering with easy editing

    [Zhang et al., 2023] Yi Zhang, Xiaoyang Huang, Bingbing Ni, Teng Li, and Wenjun Zhang. Frequency-modulated point cloud rendering with easy editing. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pat- tern Recognition, pages 119–129,

    2023

  19. [38]

    View synthesis with sculpted neural points

    [Zuo and Deng, 2022] Yiming Zuo and Jia Deng. View synthesis with sculpted neural points. arXiv preprint arXiv:2205.05869, 2022

    2022 arXiv

  20. [2009]

    Pointnerf++: A multi- scale, point-based neural radiance field,

    [Sun et al., 2023] Weiwei Sun, Eduard Trulls, Yang-Che Tseng, Sneha Sambandam, Gopal Sharma, Andrea Tagliasacchi, and Kwang Moo Yi. Pointnerf++: A multi- scale, point-based neural radiance field,

    2023

  21. [2014]

    3d gaus- sian splatting for real-time radiance field rendering

    [Kerbl et al., 2023] Bernhard Kerbl, Georgios Kopanas, Thomas Leimk ¨uhler, and George Drettakis. 3d gaus- sian splatting for real-time radiance field rendering. ACM Transactions on Graphics, 42(4), July

    2023

  22. [2015]

    Adop: Approximate differentiable one- pixel point rendering

    [R¨uckert et al., 2022] Darius R ¨uckert, Linus Franke, and Marc Stamminger. Adop: Approximate differentiable one- pixel point rendering. ACM Transactions on Graphics (ToG), 41(4):1–14,

    2022

  23. [2016]

    OpenGL program- ming guide: the official guide to learning OpenGL, ver- sions 3.0 and 3.1

    [Shreiner et al., 2009] Dave Shreiner, Bill The Khronos OpenGL ARB Working Group, et al. OpenGL program- ming guide: the official guide to learning OpenGL, ver- sions 3.0 and 3.1. Pearson Education,

    2009

  24. [2017]

    Neural point cloud ren- dering via multi-plane projection

    [Dai et al., 2020] Peng Dai, Yinda Zhang, Zhuwen Li, Shuaicheng Liu, and Bing Zeng. Neural point cloud ren- dering via multi-plane projection. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 7830–7839,

    2020

  25. [2018]

    Dai, and Quoc V

    [Ha et al., 2017] David Ha, Andrew M. Dai, and Quoc V . Le. Hypernetworks. In International Conference on Learning Representations,

    2017

  26. [2020]

    Fairchild

    [Andersson et al., 2020] Pontus Andersson, Jim Nilsson, Tomas Akenine-M ¨oller, Magnus Oskarsson, Kalle ˚Astr¨om, and Mark D. Fairchild. FLIP: A Difference Evaluator for Alternating Images. Proceedings of the ACM on Computer Graphics and Interactive Techniques , 3(2):15:1–15:23,

    2020

  27. [2021]

    U-net: Convolutional networks for biomedical image segmentation

    [Ronneberger et al., 2015] Olaf Ronneberger, Philipp Fis- cher, and Thomas Brox. U-net: Convolutional networks for biomedical image segmentation. In Medical Image Computing and Computer-Assisted Intervention–MICCAI 2015: 18th International Conference, Munich, Germany, October ...

    2015

  28. [2022]

    Scannet: Richly-annotated 3d reconstructions of indoor scenes

    [Dai et al., 2017] Angela Dai, Angel X Chang, Manolis Savva, Maciej Halber, Thomas Funkhouser, and Matthias Nießner. Scannet: Richly-annotated 3d reconstructions of indoor scenes. In Proceedings of the IEEE conference on computer vision and pattern recognition , pages 5828– 5839,

    2017

  29. [2023]

    Tensorf: Tensorial radi- ance fields

    [Chen et al., 2022] Anpei Chen, Zexiang Xu, Andreas Geiger, Jingyi Yu, and Hao Su. Tensorf: Tensorial radi- ance fields. In European Conference on Computer Vision, pages 333–350. Springer,

    2022

  30. [2024]

    Burst denoising with kernel prediction networks

    [Mildenhall et al., 2018] Ben Mildenhall, Jonathan T Bar- ron, Jiawen Chen, Dillon Sharlet, Ren Ng, and Robert Car- roll. Burst denoising with kernel prediction networks. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 2502–2510,

    2018

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.