pith. sign in

arxiv: 2605.25345 · v1 · pith:ZWDENRDQnew · submitted 2026-05-25 · 💻 cs.GR · cs.CV

Depth Peeling for High-Fidelity Gaussian-Enhanced Surfel Rendering

Keyang Ye , Hongzhi Wu , Kun Zhou This is my paper

Pith reviewed 2026-06-29 20:02 UTC · model grok-4.3

classification 💻 cs.GR cs.CV
keywords depth peelingGaussian-enhanced surfelssort-free renderingtransmittance modulationnovel view synthesisaliasing artifacts3D Gaussian splattingsurfels
0
0 comments X

The pith

Augmenting Gaussian-enhanced surfels with semi-transparent boundaries and applying depth peeling yields correct per-pixel ordering for sort-free rendering.

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

The paper introduces DP-GES by adding semi-transparent boundaries to the opaque surfels of Gaussian-Enhanced Surfels. Depth peeling then determines the precise order of these boundaries at each pixel to support accurate color blending. This produces sort-free Gaussian splatting that applies correct transmittance modulation. The changes remove aliasing and popping while allowing fully differentiable optimization of the full representation. Experiments across scenes show improved reconstruction quality over prior methods.

Core claim

DP-GES augments opaque surfels with semi-transparent boundaries and leverages Depth Peeling to establish accurate per-pixel ordering. This design enables sort-free Gaussian splatting with correct transmittance modulation, effectively eliminating aliasing and popping artifacts while facilitating a fully differentiable joint optimization.

What carries the argument

Depth Peeling applied to semi-transparent surfel boundaries to compute per-pixel ordering and transmittance

If this is right

  • Correct transmittance modulation without sorting primitives
  • Removal of aliasing and popping artifacts in rendered views
  • Support for fully differentiable joint optimization of the representation
  • Higher reconstruction quality than previous sort-free approaches across tested scenes

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The per-pixel ordering step could be adapted to other point or surfel representations that currently rely on approximate blending.
  • Differentiability may allow tighter coupling with gradient-based reconstruction pipelines that optimize both geometry and appearance.
  • Reduced popping could improve temporal stability when the method is applied to video or dynamic scene capture.

Load-bearing premise

Depth peeling on semi-transparent surfel boundaries will produce accurate per-pixel ordering and transmittance modulation without new artifacts or costs that offset the sort-free advantage.

What would settle it

Side-by-side renders of the same scene showing visible aliasing or popping artifacts in DP-GES output, or quantitative metrics such as PSNR no higher than the baseline GES method.

Figures

Figures reproduced from arXiv: 2605.25345 by Hongzhi Wu, Keyang Ye, Kun Zhou.

Figure 1
Figure 1. Figure 1: Our novel representation achieves fast, high-fidelity and popping-free rendering, outperforming state-of-the-art techniques: [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Representation and rendering pipeline of DP-GES. The DP-GES representation is composed of a set of 2D opaque surfels with [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Comparison of our surfel representation and the sur [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Qualitative comparisons on image quality. From top to bottom: [PITH_FULL_IMAGE:figures/full_fig_p006_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Qualitative comparisons with GES [42]. GES exhibits noticeable aliasing artifacts due to its depth test. Ground-truth Ours GES [PITH_FULL_IMAGE:figures/full_fig_p007_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Qualitative comparisons with GES [ [PITH_FULL_IMAGE:figures/full_fig_p007_6.png] view at source ↗
Figure 8
Figure 8. Figure 8: Qualitative comparisons of “popping” artifacts. The [PITH_FULL_IMAGE:figures/full_fig_p007_8.png] view at source ↗
Figure 9
Figure 9. Figure 9: Qualitative comparisons of rendered surfel colors with [PITH_FULL_IMAGE:figures/full_fig_p008_9.png] view at source ↗
Figure 10
Figure 10. Figure 10: Qualitative comparisons of images rendered without [PITH_FULL_IMAGE:figures/full_fig_p008_10.png] view at source ↗
Figure 11
Figure 11. Figure 11: Qualitative comparisons of images rendered with [PITH_FULL_IMAGE:figures/full_fig_p008_11.png] view at source ↗
Figure 12
Figure 12. Figure 12: Qualitative comparisons of images rendered with the [PITH_FULL_IMAGE:figures/full_fig_p011_12.png] view at source ↗
Figure 13
Figure 13. Figure 13: Qualitative comparisons of images rendered without [PITH_FULL_IMAGE:figures/full_fig_p012_13.png] view at source ↗
Figure 14
Figure 14. Figure 14: Qualitative comparisons between our DP-GES and GES [ [PITH_FULL_IMAGE:figures/full_fig_p013_14.png] view at source ↗
read the original abstract

Novel view synthesis has been significantly advanced by NeRFs and 3D Gaussian Splatting (3DGS), which require ordering volumetric samples or primitives for correct color blending. While the recent Gaussian-Enhanced Surfels (GES) enable high-performance, sort-free rendering, they suffer from aliasing artifacts and suboptimal reconstruction. To address these limitations, we propose DP-GES, a novel representation that augments opaque surfels with semi-transparent boundaries and leverages Depth Peeling to establish accurate per-pixel ordering. This design enables sort-free Gaussian splatting with correct transmittance modulation, effectively eliminating aliasing and popping artifacts while facilitating a fully differentiable joint optimization. Extensive experiments demonstrate that our method achieves superior reconstruction quality and compares favorably against state-of-the-art techniques across a wide range of scenes.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

0 major / 0 minor

Summary. The paper proposes DP-GES, a novel representation that augments Gaussian-Enhanced Surfels (GES) with semi-transparent boundaries and applies depth peeling to obtain accurate per-pixel ordering. This enables sort-free Gaussian splatting with correct transmittance modulation, claimed to eliminate aliasing and popping artifacts while supporting fully differentiable joint optimization. Extensive experiments are said to show superior reconstruction quality relative to state-of-the-art methods across diverse scenes.

Significance. If the depth-peeling construction delivers the claimed ordering and transmittance accuracy without offsetting costs, the work would meaningfully advance high-performance novel-view synthesis by reconciling the performance benefits of sort-free surfel methods with the blending correctness previously requiring explicit sorting or volumetric ordering.

Simulated Author's Rebuttal

0 responses · 0 unresolved

We thank the referee for their review of our manuscript on DP-GES. We appreciate the positive framing of the work's potential significance in reconciling sort-free rendering with correct blending. No specific major comments were listed in the report, so we provide no point-by-point responses below.

Circularity Check

0 steps flagged

No significant circularity detected

full rationale

The abstract and description contain no equations, derivations, self-citations, or load-bearing steps that reduce predictions or results to fitted inputs by construction. The method is presented at a conceptual level without any claimed derivation chain, uniqueness theorems, or ansatzes that could be inspected for circularity. This matches the default case of a self-contained high-level description with no identifiable reduction to its own inputs.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Abstract-only review supplies no identifiable free parameters, axioms, or invented entities.

pith-pipeline@v0.9.1-grok · 5658 in / 1136 out tokens · 37613 ms · 2026-06-29T20:02:13.794336+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

51 extracted references · 1 linked inside Pith

  1. [1]

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

    Jonathan T Barron, Ben Mildenhall, Dor Verbin, Pratul P Srinivasan, and Peter Hedman. Mip-nerf 360: Unbounded anti-aliased neural radiance fields. InProceedings of the IEEE/CVF CVPR, pages 5470–5479, 2022. 5, 7

    2022

  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. In2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 5460–5469, 2022. 2, 8

    2022

  3. [3]

    Textured gaussians for en- hanced 3d scene appearance modeling

    Brian Chao, Hung-Yu Tseng, Lorenzo Porzi, Chen Gao, Tuotuo Li, Qinbo Li, Ayush Saraf, Jia-Bin Huang, Johannes Kopf, Gordon Wetzstein, et al. Textured gaussians for en- hanced 3d scene appearance modeling. InProceedings of the Computer Vision and Pattern Recognition Conference, pages 8964–8974, 2025. 2

    2025

  4. [4]

    A survey on 3d gaussian splatting, 2024

    Guikun Chen and Wenguan Wang. A survey on 3d gaussian splatting, 2024. 2

    2024

  5. [5]

    Be- yond gaussians: Fast and high-fidelity 3d splatting with lin- ear kernels, 2024

    Haodong Chen, Runnan Chen, Qiang Qu, Zhaoqing Wang, Tongliang Liu, Xiaoming Chen, and Yuk Ying Chung. Be- yond gaussians: Fast and high-fidelity 3d splatting with lin- ear kernels, 2024. 2

    2024

  6. [6]

    Interactive order-independent transparency

    Cass Everitt. Interactive order-independent transparency. White paper, nVIDIA, 2(6):7, 2001. 3

    2001

  7. [7]

    Nerf: Neural radiance field in 3d vision, a comprehensive review.arXiv preprint arXiv:2210.00379,

    Kyle Gao, Yina Gao, Hongjie He, Dening Lu, Linlin Xu, and Jonathan Li. Nerf: Neural radiance field in 3d vision, a comprehensive review.arXiv preprint arXiv:2210.00379,

    Pith/arXiv arXiv

  8. [8]

    Gortler, Radek Grzeszczuk, Richard Szeliski, and Michael F

    Steven J. Gortler, Radek Grzeszczuk, Richard Szeliski, and Michael F. Cohen. The lumigraph. InProceedings of SIG- GRAPH ’96, page 43–54, New York, NY , USA, 1996. ACM. 1

    1996

  9. [9]

    The accumulation buffer: hardware support for high-quality rendering.SIGGRAPH Comput

    Paul Haeberli and Kurt Akeley. The accumulation buffer: hardware support for high-quality rendering.SIGGRAPH Comput. Graph., 24(4):309–318, 1990. 3

    1990

  10. [10]

    Efficient perspective-correct 3d gaussian splatting using hybrid trans- parency, 2024

    Florian Hahlbohm, Fabian Friederichs, Tim Weyrich, Li- nus Franke, Moritz Kappel, Susana Castillo, Marc Stam- minger, Martin Eisemann, and Marcus Magnor. Efficient perspective-correct 3d gaussian splatting using hybrid trans- parency, 2024. 2

    2024

  11. [11]

    Efficient perspective-correct 3d gaussian splatting using hybrid trans- parency.Computer Graphics Forum, 44(2), 2025

    Florian Hahlbohm, Fabian Friederichs, Tim Weyrich, Li- nus Franke, Moritz Kappel, Susana Castillo, Marc Stam- minger, Martin Eisemann, and Marcus Magnor. Efficient perspective-correct 3d gaussian splatting using hybrid trans- parency.Computer Graphics Forum, 44(2), 2025. 1, 5, 7

    2025

  12. [12]

    Ges: Generalized exponential splatting for efficient radiance field rendering

    Abdullah Hamdi, Luke Melas-Kyriazi, Jinjie Mai, Guocheng Qian, Ruoshi Liu, Carl V ondrick, Bernard Ghanem, and Andrea Vedaldi. Ges: Generalized exponential splatting for efficient radiance field rendering. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 19812–19822, 2024. 2

    2024

  13. [13]

    Deep blending for free-viewpoint image-based rendering.ACM Trans

    Peter Hedman, Julien Philip, True Price, Jan-Michael Frahm, George Drettakis, and Gabriel Brostow. Deep blending for free-viewpoint image-based rendering.ACM Trans. Graph., 37(6):1–15, 2018. 5

    2018

  14. [14]

    3d convex splatting: Radiance field rendering with 3d smooth convexes

    Jan Held, Renaud Vandeghen, Abdullah Hamdi, Adrien Deliege, Anthony Cioppa, Silvio Giancola, Andrea Vedaldi, Bernard Ghanem, and Marc Van Droogenbroeck. 3d convex splatting: Radiance field rendering with 3d smooth convexes. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2025. 2

    2025

  15. [15]

    Sort-free gaussian splatting via weighted sum rendering

    Qiqi Hou, Randall Rauwendaal, Zifeng Li, Hoang Le, Farzad Farhadzadeh, Fatih Porikli, Alexei Bourd, and Amir Said. Sort-free gaussian splatting via weighted sum rendering. arXiv preprint arXiv:2410.18931, 2024. 5

    arXiv 2024

  16. [16]

    Sort-free gaussian splatting via weighted sum rendering,

    Qiqi Hou, Randall Rauwendaal, Zifeng Li, Hoang Le, Farzad Farhadzadeh, Fatih Porikli, Alexei Bourd, and Amir Said. Sort-free gaussian splatting via weighted sum rendering,

  17. [17]

    2d gaussian splatting for geometrically ac- curate radiance fields

    Binbin Huang, Zehao Yu, Anpei Chen, Andreas Geiger, and Shenghua Gao. 2d gaussian splatting for geometrically ac- curate radiance fields. InACM SIGGRAPH 2024 Conference Papers, 2024. 2, 8, 1

    2024

  18. [18]

    Deformable radial kernel splatting

    Yi-Hua Huang, Ming-Xian Lin, Yang-Tian Sun, Ziyi Yang, Xiaoyang Lyu, Yan-Pei Cao, and Xiaojuan Qi. Deformable radial kernel splatting. InProceedings of the Computer Vision and Pattern Recognition Conference, pages 21513– 21523, 2025. 2

    2025

  19. [19]

    3d gaussian splatting for real-time radiance field rendering.ACM Trans

    Bernhard Kerbl, Georgios Kopanas, Thomas Leimk ¨uhler, and George Drettakis. 3d gaussian splatting for real-time radiance field rendering.ACM Trans. Graph., 42(4), 2023. 1, 2, 3, 5

    2023

  20. [20]

    3d gaussian splat- ting as markov chain monte carlo.Advances in Neural In- formation Processing Systems, 37:80965–80986, 2024

    Shakiba Kheradmand, Daniel Rebain, Gopal Sharma, Wei- wei Sun, Yang-Che Tseng, Hossam Isack, Abhishek Kar, Andrea Tagliasacchi, and Kwang Moo Yi. 3d gaussian splat- ting as markov chain monte carlo.Advances in Neural In- formation Processing Systems, 37:80965–80986, 2024. 1, 2, 5

    2024

  21. [21]

    Stochasticsplats: Stochastic rasteriza- tion for sorting-free 3d gaussian splatting.arXiv preprint arXiv:2503.24366, 2025

    Shakiba Kheradmand, Delio Vicini, George Kopanas, Dmitry Lagun, Kwang Moo Yi, Mark Matthews, and An- drea Tagliasacchi. Stochasticsplats: Stochastic rasteriza- tion for sorting-free 3d gaussian splatting.arXiv preprint arXiv:2503.24366, 2025. 2

    arXiv 2025

  22. [22]

    Tanks and temples: Benchmarking large-scale scene reconstruction.ACM Trans

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

    2017

  23. [23]

    Light field rendering

    Marc Levoy and Pat Hanrahan. Light field rendering. In Proceedings of SIGGRAPH ’96, page 31–42, New York, NY , USA, 1996. Association for Computing Machinery. 1

    1996

  24. [24]

    A survey on neural radiance fields

    Yun Liao, Yide Di, Hao Zhou, Kaijun Zhu, Mingyu Lu, Qing Duan, and Junhui Liu. A survey on neural radiance fields. ACM Comput. Surv., 58(2), 2025. 2

    2025

  25. [25]

    Deformable beta splatting

    Rong Liu, Dylan Sun, Meida Chen, Yue Wang, and Andrew Feng. Deformable beta splatting. InProceedings of the Spe- cial Interest Group on Computer Graphics and Interactive Techniques Conference Conference Papers, New York, NY , USA, 2025. Association for Computing Machinery. 1, 2, 3, 4, 5, 8

    2025

  26. [26]

    Duplex-gs: Proxy-guided weighted blending for real- time order-independent gaussian splatting.arXiv preprint arXiv:2508.03180, 2025

    Weihang Liu, Yuke Li, Yuxuan Li, Jingyi Yu, and Xin Lou. Duplex-gs: Proxy-guided weighted blending for real- time order-independent gaussian splatting.arXiv preprint arXiv:2508.03180, 2025. 3

    arXiv 2025

  27. [27]

    Ever: Exact volu- metric ellipsoid rendering for real-time view synthesis

    Alexander Mai, Peter Hedman, George Kopanas, Dor Verbin, David Futschik, Qiangeng Xu, Falko Kuester, Jonathan T Barron, and Yinda Zhang. Ever: Exact volu- metric ellipsoid rendering for real-time view synthesis. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 4930–4939, 2025. 2

    2025

  28. [28]

    Weighted blended order- independent transparency.Journal of Computer Graphics Techniques, 2(4), 2013

    Morgan McGuire and Louis Bavoil. Weighted blended order- independent transparency.Journal of Computer Graphics Techniques, 2(4), 2013. 3

    2013

  29. [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. InProceedings of ECCV, pages 405–421, 2020. 1, 5, 3

    2020

  30. [30]

    3d gaussian ray trac- ing: Fast tracing of particle scenes.ACM Trans

    Nicolas Moenne-Loccoz, Ashkan Mirzaei, Or Perel, Ric- cardo de Lutio, Janick Martinez Esturo, Gavriel State, Sanja Fidler, Nicholas Sharp, and Zan Gojcic. 3d gaussian ray trac- ing: Fast tracing of particle scenes.ACM Trans. Graph., 43 (6):1–19, 2024. 2

    2024

  31. [31]

    Gaussian-plus-sdf slam: High-fidelity 3d reconstruction at 150+ fps.Computa- tional Visual Media, 2025

    Zhexi Peng, Kun Zhou, and Tianjia Shao. Gaussian-plus-sdf slam: High-fidelity 3d reconstruction at 150+ fps.Computa- tional Visual Media, 2025. 1

    2025

  32. [32]

    Stopthepop: Sorted gaussian splatting for view-consistent real-time rendering.ACM Trans

    Lukas Radl, Michael Steiner, Mathias Parger, Alexan- der Weinrauch, Bernhard Kerbl, and Markus Steinberger. Stopthepop: Sorted gaussian splatting for view-consistent real-time rendering.ACM Trans. Graph., 43(4), 2024. 2, 5, 7

    2024

  33. [33]

    Sch ¨onberger and Jan-Michael Frahm

    Johannes L. Sch ¨onberger and Jan-Michael Frahm. Structure- from-motion revisited. In2016 IEEE Conference on Com- puter Vision and Pattern Recognition (CVPR), pages 4104– 4113, 2016. 5

    2016

  34. [34]

    Layered depth images

    Jonathan Shade, Steven Gortler, Li-wei He, and Richard Szeliski. Layered depth images. InProceedings of the 25th annual conference on Computer graphics and interac- tive techniques, pages 231–242, 1998. 2, 3

    1998

  35. [35]

    Solidgs: Consolidating gaussian surfel splatting for sparse-view sur- face reconstruction.arXiv preprint arXiv:2412.15400, 2024

    Zhuowen Shen, Yuan Liu, Zhang Chen, Zhong Li, Jiepeng Wang, Yongqing Liang, Zhengming Yu, Jingdong Zhang, Yi Xu, Scott Schaefer, Xin Li, and Wenping Wang. Solidgs: Consolidating gaussian surfel splatting for sparse-view sur- face reconstruction.arXiv preprint arXiv:2412.15400, 2024. 2

    arXiv 2024

  36. [36]

    Hdgs: Textured 2d gaussian splatting for enhanced scene rendering.arXiv preprint arXiv:2412.01823,

    Yunzhou Song, Heguang Lin, Jiahui Lei, Lingjie Liu, and Kostas Daniilidis. Hdgs: Textured 2d gaussian splatting for enhanced scene rendering.arXiv preprint arXiv:2412.01823,

    arXiv

  37. [37]

    Billboard splatting (bbsplat): Learnable textured primitives for novel view synthesis

    David Svitov, Pietro Morerio, Lourdes Agapito, and Alessio Del Bue. Billboard splatting (bbsplat): Learnable textured primitives for novel view synthesis. InProceedings of the IEEE/CVF International Conference on Computer Vision, pages 25029–25039, 2025. 2

    2025

  38. [39]

    Recent advances in 3d gaussian splatting.Computational Visual Media, 10(4):613– 642, 2024

    Tong Wu, Yu-Jie Yuan, Ling-Xiao Zhang, Jie Yang, Yan- Pei Cao, Ling-Qi Yan, and Lin Gao. Recent advances in 3d gaussian splatting.Computational Visual Media, 10(4):613– 642, 2024. 2

    2024

  39. [40]

    Spec-gaussian: Anisotropic view-dependent ap- pearance for 3d gaussian splatting.Advances in Neural In- formation Processing Systems, 37:61192–61216, 2024

    Ziyi Yang, Xinyu Gao, Yang-Tian Sun, Yihua Huang, Xi- aoyang Lyu, Wen Zhou, Shaohui Jiao, Xiaojuan Qi, and Xi- aogang Jin. Spec-gaussian: Anisotropic view-dependent ap- pearance for 3d gaussian splatting.Advances in Neural In- formation Processing Systems, 37:61192–61216, 2024. 2

    2024

  40. [41]

    3d gaussian splat- ting with deferred reflection

    Keyang Ye, Qiming Hou, and Kun Zhou. 3d gaussian splat- ting with deferred reflection. InACM SIGGRAPH 2024 Con- ference Papers, 2024. 2

    2024

  41. [42]

    When gaussian meets surfel: Ultra-fast high-fidelity radiance field render- ing.ACM Trans

    Keyang Ye, Tianjia Shao, and Kun Zhou. When gaussian meets surfel: Ultra-fast high-fidelity radiance field render- ing.ACM Trans. Graph., 44(4), 2025. 1, 3, 4, 5, 7, 2

    2025

  42. [43]

    Absgs: Recovering fine details in 3d gaussian splat- ting

    Zongxin Ye, Wenyu Li, Sidun Liu, Peng Qiao, and Yong Dou. Absgs: Recovering fine details in 3d gaussian splat- ting. InProceedings of the 32nd ACM International Confer- ence on Multimedia, pages 1053–1061, 2024. 1, 2, 5, 7

    2024

  43. [44]

    Mip-splatting: Alias-free 3d gaussian splat- ting

    Zehao Yu, Anpei Chen, Binbin Huang, Torsten Sattler, and Andreas Geiger. Mip-splatting: Alias-free 3d gaussian splat- ting. In2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 19447–19456, 2024. 5

    2024

  44. [45]

    Neural shell texture splatting: More details and fewer primitives

    Xin Zhang, Anpei Chen, Jincheng Xiong, Pinxuan Dai, Yu- jun Shen, and Weiwei Xu. Neural shell texture splatting: More details and fewer primitives. InProceedings of the IEEE/CVF International Conference on Computer Vision, pages 25229–25238, 2025. 2

    2025

  45. [46]

    3d stu- dent splatting and scooping

    Jialin Zhu, Jiangbei Yue, Feixiang He, and He Wang. 3d stu- dent splatting and scooping. InProceedings of the Computer Vision and Pattern Recognition Conference (CVPR), pages 21045–21054, 2025. 1, 2, 5, 7 Depth Peeling for High-Fidelity Gaussian-Enhanced Surfel Rendering Supplementary Material In the supplementary material, Sec. 8–Sec. 11 describe more te...

    2025

  46. [47]

    4, we use the depth marginϵ s to prevent Gaussians intersecting with surfels from being partially truncated dur- ing the depth test

    Determining Depth Margin In Sec. 4, we use the depth marginϵ s to prevent Gaussians intersecting with surfels from being partially truncated dur- ing the depth test. Specifically, we first compute the initial value ofϵ i = 5 2 (sX i +s Y i )in the same way as GES, and then, for each surfel, we query its 16 nearest neighboring surfels and take the median o...

  47. [48]

    Specifically, for surfels, we prune surfels that are either heavily occluded or extremely small, determined by the number of pixels they cover in the first peeled layer

    Training Details Joint Optimization.During this stage, we adopt the den- sification and pruning strategies of GES with minor modi- fications. Specifically, for surfels, we prune surfels that are either heavily occluded or extremely small, determined by the number of pixels they cover in the first peeled layer. The pruning thresholds are the same as GES. F...

    2000

  48. [49]

    OpenGL Implementation As demonstrated in Sec. 5, we develop an OpenGL-based renderer for real-time rendering after training, fully leverag- ing the efficiency of our sort-free rendering scheme within the traditional graphics pipeline. The OpenGL-based ren- derer also consists of two main passes: In the first pass, we perform depth peeling for the nearest ...

  49. [50]

    Geometry Regularization To enable the surfel representation to better capture the scene geometry, which is essential for downstream tasks such as mesh extraction, we optionally adopt a geometry regularization lossL geo =λ 4Lsd +λ5Lsn +λ6Lsdn, where λ4 = 0.1,λ 5 = 0.05andλ 6 = 0.05, respectively. Lsd =L 1(Dsupv, Ds),(6) Lsn = 1 HW X ˆx (1−N supv(ˆx)·N s(ˆx...

  50. [51]

    14, we visualize the surfel-Gaussian decomposition rendering compared with GES

    Additional Comparisons with GES Surfel-Gaussian Decomposition Rendering.In the left- most 3 columns of Fig. 14, we visualize the surfel-Gaussian decomposition rendering compared with GES. InChair, GES preserves high-frequency texture in surfel colors be- cause its image-loss-based surfel initialization inevitably embeds appearance details, deviating from ...

  51. [52]

    The additive study demonstrates a clear and continuous improvement as we progressively upgrade the original GES to our full model

    Additional Ablations We present quantitative additive ablations and additional re- sults in Table 4. The additive study demonstrates a clear and continuous improvement as we progressively upgrade the original GES to our full model. Applying our render- ing model to GES while disabling the new gradient path (Base, w/o trans. grad) alleviates the aliasing i...