Pith. sign in

REVIEW 3 major objections 6 minor 3 cited by

Gaussian Billboards: Expressive 2D Gaussian Splatting with Textures

T0 review · 3 major / 6 minor · reviewed 2026-08-11 · deepseek-v4-flash

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

Pith's one-line read Giving each 2D Gaussian splat a tiny color texture sharpens reconstruction without adding primitives.

desk verdict A clean, honestly reported incremental improvement to 2DGS that replaces per-splat solid colors with small texture grids; the evidence is consistent if not over-strong, and the paper deserves a serious referee. read the letter →

arxiv 2412.12734 v1 pith:Z5W6PV7R submitted 2024-12-17 cs.CV cs.GR

Sebastian Weiss , Derek Bradley This is my paper · ORCID
classification cs.CVcs.GR
keywords 2DGaussianSplattingBillboardsper-primitivetexturebilinearinterpolationnovelviewsynthesisscenereconstructionmapping
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

Gaussian Billboards is an extension of 2D Gaussian Splatting that gives each flat, oriented splat its own small color texture instead of a single solid color. The paper argues that because most scenes carry more texture information than geometric information, a per-splat texture lets the same or fewer primitives reconstruct sharp detail, especially on flat but richly textured surfaces such as posters, marble, foliage, and faces. The authors show in overfitting experiments that texture resolution N=4 with spatial extent σ=0.5 improves PSNR, SSIM, and LPIPS, and that on NeRF360 scenes and face captures the textured splats improve visual quality without changing training hyperparameters. A sympathetic reader would take the central claim to be: spatially varying per-splat color is a simple, orthogonal upgrade to 2DGS that improves reconstruction sharpness and quality.

What carries the argument

The central object is the per-primitive texture grid, an N×N array of RGB colors parameterized by the same (u,v) tangent-plane coordinates that already define each 2D Gaussian splat. During rasterization the pixel color of a primitive is computed by bilinear interpolation with border clamping: c(u,v)=bilinear(C, u/σ, v/σ), so the texture rotates and scales with the splat's tangent vectors t_u, t_v and scales s_u, s_v. The machinery works because gradients flow through the interpolation to the grid and to the uv coordinates, hence to position, scale, and orientation, and the fused CUDA kernels of gsplat carry the extra per-primitive parameters in shared memory. View-dependent appearance is handled by a diffuse-specular split: the grid stores the Lambertian albedo and a single per-primitive set of spherical harmonics coefficients adds a view-dependent residual.

What would settle it

Train Gaussian Billboards and baseline 2DGS on a scene with a large flat textured surface observed by many training cameras, then evaluate on several held-out cameras over multiple random initializations; if the novel-view SSIM or LPIPS advantage over 2DGS disappears, or if the learned color grids contain view-dependent structure visible as per-camera patterns, the central claim fails.

Watch

Extended reading notes

Core claim

On the paper's own terms, the discovery is that replacing the constant color c_k of each 2D Gaussian primitive with a spatially varying color obtained by bilinearly interpolating an N×N color grid C_k in the primitive's uv-frame improves reconstruction capability both when the number of primitives is fixed and when densification and pruning adapt the count during optimization. The color at a point is c(u,v)_k = bilinear(C_k, u/σ, v/σ), with texture coordinates clamped to [−σ, +σ]^2; σ=0.5 places most of the texture inside the high-opacity core of the Gaussian. The paper reports that this modification yields consistently better LPIPS on NeRF360 validation images, better SSIM in five of seven scenes, sharper flat textured regions, and, on a face capture, similar metrics with roughly 64,000 primitives instead of 90,000. The authors frame the result as merging classic billboard texturing with modern splatting optimization.

Load-bearing premise

The load-bearing premise is that a small per-splat texture, optimized from scratch with only a photometric loss and no texture-specific regularization, will remain view-consistent rather than memorizing per-view color patterns in the local uv frame; if it memorizes, the sharpness gains would not transfer to novel views.

Editorial extensions

If this is right

  • With a fixed primitive budget, e.g. 10,000 or 100,000 splats, image-fitting PSNR, SSIM, and LPIPS improve monotonically from N=1 (solid color) up to N=8, with N=4 recommended for the best quality-to-cost trade-off.
  • Under standard 2DGS densification and pruning with unchanged hyperparameters, Gaussian Billboards improve novel-view quality on NeRF360 scenes, most visibly on flat textured regions like marble counters and vegetation.
  • On a multi-view face capture, the textured representation reaches comparable metrics with fewer primitives (about 64,000 versus 90,000) and a slightly faster test-view render.
  • The diffuse-specular split keeps the texture view-independent, so the method remains compatible with other 2DGS and 3DGS improvements as an orthogonal change.

Reading between the lines

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

  • If small per-splat textures remain view-consistent under a purely photometric loss, the same idea could be applied to other per-primitive attributes such as opacity or normal maps, potentially sharpening geometry as well as color.
  • The shared-memory ceiling that currently limits N to 8 suggests a testable extension: storing textures in a compressed or hashed form, or sharing a small atlas across splats, could lift the resolution limit without the large training slowdown at N=8.
  • The texture grid decouples appearance from geometry more cleanly than solid colors, so texture-space editing or relighting of individual splats becomes a natural next step; the paper does not explore this.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 6 minor

Summary. The paper introduces Gaussian Billboards, an extension of 2D Gaussian Splatting (2DGS) in which each 2D primitive is assigned a small N×N texture grid (default 4×4) whose colors are bilinearly interpolated in the primitive's local uv-plane, replacing the per-primitive solid color. The method is trained end-to-end with the same photometric loss and densification/pruning scheme as 2DGS, and is integrated into the gsplat CUDA kernels with manually derived adjoint code. The authors evaluate on single-image overfitting across three images with ablations of the texture resolution N and the spatial extent parameter σ, and on 3D scene reconstruction using the NeRF360 dataset and a multi-view face capture with one held-out test camera. They report consistent LPIPS improvements and visual sharpness gains, with N=4 and σ=0.5 selected as the recommended hyperparameters, and they discuss the relationship to the concurrent GStex method.

Significance. If the central claim holds, the paper is a simple yet effective contribution: it decouples geometric and texture capacity in 2DGS, improving reconstruction of textured flat regions while using equal or fewer primitives. The strengths are the clean formulation (Eq. 5), the detailed ablation of N and σ, the transparent integration details and adjoint derivations in the supplementary material, and the qualitative demonstration on challenging scenes. The main weakness is that the novel-view evidence for the central claim is limited to near-training views and a single held-out camera, leaving the possibility of per-splat texture overfitting unresolved. The absence of error bars and the use of the same metrics for hyperparameter selection further weaken the quantitative claim of consistent improvement. Overall, this is a worthwhile enhancement that needs a more rigorous evaluation of view consistency and statistical significance.

major comments (3)
  1. [Sec. 4.2, Figs. 7-8] The only genuinely held-out novel-view evidence is a single central test camera for the face reconstruction (Fig. 7), while the NeRF360 validation images are typically sampled from the same camera trajectory as the training frames and therefore have limited baseline. Since the per-splat textures are optimized with only a photometric loss and no cross-splat or cross-view consistency constraint (Sec. 3.3), the reported LPIPS gains could in principle come from view-dependent patterns encoded in the texture grids rather than improved scene appearance; without a large-baseline held-out evaluation or a cross-view texture-consistency check, the central claim of improved novel view synthesis is not fully supported.
  2. [Sec. 4.1, Figs. 5-6] The recommended hyperparameters N=4 and σ=0.5 are selected by measuring exactly the same SSIM and LPIPS metrics on which the final comparisons are made, and the ablation uses only three single-image fits with no repeated runs. Given that several final numbers in Fig. 8 show PSNR values slightly below the baseline and SSIM differences below 0.001, the phrase 'consistently outperforms' is too strong without error bars or a separate validation set for hyperparameter selection.
  3. [Sec. 3.3, Eq. (5)] The texture grid of each primitive is initialized with slight color variations and trained without any regularization that would encourage agreement between overlapping primitives or stability under viewpoint changes. When the underlying geometry is imperfect, the optimizer has substantial freedom to store view-dependent shading in the texture, which is a concrete mechanism for the overfitting concern; the paper should include an analysis or experiment demonstrating that the learned texture grids correspond to stable surface appearance across views.
minor comments (6)
  1. [Figs. 4 and 8] The acronym PSNR is misspelled as 'PNSR' in several captions and text lines; please correct these typographical errors.
  2. [Fig. 5] The caption lists only σ values 0.25, 0.5, 1.0, and 2.0, while the x-axis extends to 4.0 and 0.125; please align the caption with the plotted range.
  3. [References] Reference [21] has a stray '1' at the end of the page field; please fix the bibliographic entry.
  4. [Sec. 3.2] The phrase 'the full color grid residing within the isoline of 0.8 opacity' is unclear; consider rewording to 'the texture square [−σ,σ]² approximately coincides with the 0.8-opacity isoline of the Gaussian'.
  5. [Sec. 3.3] The paper claims the method is orthogonal to other improvements but does not provide a runtime or memory overhead comparison in absolute terms; a short paragraph quantifying the added parameters per splat (e.g., 48 RGB values for N=4) would help readers gauge the cost.
  6. [Sec. 4.2] The diffuse-specular split is introduced without an ablation; consider adding a sentence justifying the use of a single SH set per primitive versus per-texel SH coefficients.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the per-splat texture is a new parameterization choice evaluated empirically; no claim reduces to a fitted parameter or to a self-citation chain.

full rationale

The paper's central claim is that replacing the per-primitive solid color of 2DGS with a small per-primitive texture improves reconstruction quality. This is supported empirically by comparisons against the 2DGS baseline on image overfitting, the NeRF360 validation set, and a face capture with a held-out test camera. Equation (5) simply defines the texture interpolation; it is not used to derive the improvement, and no fitted parameter is later relabeled as a prediction. The hyperparameters N and sigma are chosen in Sec. 4.1 via ablations on three single images, then fixed for the 3D experiments; although the same metrics (SSIM, LPIPS) appear in both selection and evaluation, the selection set is disjoint from the NeRF360 and face evaluation scenes, making this ordinary hyperparameter tuning rather than a statistically forced prediction. The paper also relies on no load-bearing self-citations: the method builds on the external 2DGS and gsplat implementations, and the concurrent GStex work is acknowledged as related rather than used to justify the approach. The limitations noted in the paper — training with only a photometric loss (Sec. 3.3) and a single held-out face camera (Sec. 4.2) — are generalization and overfitting risks, not circularity, because nothing in the equations or citations makes the novel-view improvement true by construction.

Assumptions & free parameters 2 free parameters · 5 assumptions · 0 invented entities

The central result rests on inherited 2DGS machinery and on two hand-selected hyperparameters, sigma and N. The only invented object is the per-splat texture grid, which is an optimization variable, not a new physical or geometric entity. The main unstated premises are about view consistency and the relative richness of texture versus geometry in real scenes.

free parameters (2)
  • sigma (spatial extent of texture grid) = 0.5
    Chosen via ablation on two image fitting tasks (Creation of Adam and Meteora, Fig. 5) as optimal in the range 0.25 to 4.0; used for all 3D results. It controls how much of the uv plane is covered by the texture and where border clamping begins.
  • N (texture resolution) = 4
    Recommended after ablation on three images. N=8 improves quality further but increases training time by roughly 4x due to reduced tile size, and N=4 is used for all 3D scene reconstruction results. This is a hyperparameter chosen by hand, not fitted per scene.
assumptions (5)
  • standard math The differentiable Gaussian splatting equations (Eq. 1-4) from 3DGS and 2DGS correctly model scene appearance and are used as the optimization base.
    Sec. 3.1 adopts the projection and alpha-compositing models of Kerbl et al. [9] and Huang et al. [6] without re-derivation.
  • domain assumption Each 2D primitive has a stable local uv parametrization (Eq. 3) that can carry a texture consistently across views.
    Sec. 3.2 attaches the texture grid to the primitive's tangent plane defined by tu, tv, su, sv; if this tangent parameterization drifts or is view-inconsistent, spatial textures would cause artifacts.
  • domain assumption Texture information is generally higher than geometry information in typical scenes, so per-splat texture capacity yields better reconstruction than more splats.
    Sec. 3.2 states that 'in most scenes, the texture information is higher than geometry information'; the method's benefit depends on this being true for the evaluated scenes.
  • domain assumption Optimization from photometric loss alone, without texture-specific regularization, yields view-consistent textures.
    Sec. 3.3 trains only with a photometric loss and standard 2DGS hyperparameters; no regularization prevents per-splat textures from overfitting individual training views, aside from the small grid size and bilinear interpolation.
  • domain assumption Bilinear interpolation with border clamping inside [-sigma, +sigma] is a sufficient color model for the visible part of each splat.
    Sec. 3.2 and Algorithm 2 define the color as clamped bilinear interpolation; textures outside this box or with higher frequency than the grid are not representable.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Gaussian Billboards: Expressive 2D Gaussian Splatting with Textures." pith.science (2026). https://pith.science/paper/Z5W6PV7R

@misc{pith2026241212734,
  author       = {Pith},
  title        = {Pith review of: Gaussian Billboards: Expressive 2D Gaussian Splatting with Textures},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/Z5W6PV7R}},
  note         = {Machine review of arXiv:2412.12734}
}
read the original abstract

Gaussian Splatting has recently emerged as the go-to representation for reconstructing and rendering 3D scenes. The transition from 3D to 2D Gaussian primitives has further improved multi-view consistency and surface reconstruction accuracy. In this work we highlight the similarity between 2D Gaussian Splatting (2DGS) and billboards from traditional computer graphics. Both use flat semi-transparent 2D geometry that is positioned, oriented and scaled in 3D space. However 2DGS uses a solid color per splat and an opacity modulated by a Gaussian distribution, where billboards are more expressive, modulating the color with a uv-parameterized texture. We propose to unify these concepts by presenting Gaussian Billboards, a modification of 2DGS to add spatially-varying color achieved using per-splat texture interpolation. The result is a mixture of the two representations, which benefits from both the robust scene optimization power of 2DGS and the expressiveness of texture mapping. We show that our method can improve the sharpness and quality of the scene representation in a wide range of qualitative and quantitative evaluations compared to the original 2DGS implementation.

Figures

Figures reproduced from arXiv: 2412.12734 by the authors.

Figure 1
Figure 1. We propose an enhancement for 2D Gaussian Splatting that (a) replaces the per-primitive solid color with a small per-primitive [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. The color grid of Gaussian Billboards spans the [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. The three images used for image fitting tests and hyper [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figures from the paper (5 more)
Figure 4
Figure 4. Figure 4: Overfitting a single image (a) with a fixed number of [PITH_FULL_IMAGE:figures/full_fig_p004_4.png]
Figure 5
Figure 5. Figure 5: Ablation of the spatial extent of the grid in the [PITH_FULL_IMAGE:figures/full_fig_p005_5.png]
Figure 6
Figure 6. Figure 6: Evaluation of the texture resolution N for image fitting with a fixed number of primitives, per scene with 10,000 primitives (top row) and 100,000 primitives (bottom row). With an increase in texture resolution, the Gaussian Billboards consistently lead to better fitti…
Figure 7
Figure 7. Figure 7: 3D Reconstruction of a human face from 14 training [PITH_FULL_IMAGE:figures/full_fig_p007_7.png]
Figure 8
Figure 8. Figure 8: Evaluation of traditional 2DGS compared to Gaussian Billboards ( [PITH_FULL_IMAGE:figures/full_fig_p008_8.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 3 Pith papers

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

  1. Floating Radiance Networks

    cs.CV 2026-08 conditional novelty 6.0 of 10

    FlaRe combines per-primitive latent radiance descriptors on planar Gaussians with a shared decoder and hardware ray tracing, making rendering, secondary rays, editing, and mesh extraction work in one scene model.

  2. A$^2$TG: Adaptive Anisotropic Textured Gaussians for Efficient 3D Scene Representation

    cs.CV 2026-01 conditional novelty 6.0 of 10

    Adaptive anisotropic texture allocation reduces memory in textured Gaussian splatting while keeping rendering quality competitive.

  3. Quadratic Gaussian Splatting: High Quality Surface Reconstruction with Second-order Geometric Primitives

    cs.CV 2024-11 reject novelty 6.0 of 10

    A Gaussian splatting method that fits curved paraboloid patches instead of flat disks reports better surface reconstruction, but its geodesic-distance justification is only exact for surfaces of revolution.

Reference graph

Works this paper leans on

29 extracted references · 28 canonical work pages · cited by 3 Pith papers

  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. arXiv [cs.CV], 2021. 4, 5, 8

    work page 2021

  2. [2]

    Evaluation of traditional 2DGS compared to Gaussian Billboards ( Ours) on the NeRF360 validation images [1]

    Thabo Beeler, Bernd Bickel, Paul Beardsley, Bob Sumner, PNSR=23.70, SSIM=0.653, LPIPS=0.373 Traditional 2DGS PNSR=23.62, SSIM= 0.666, LPIPS= 0.302 Ours, N = 4 PNSR=31.40, SSIM=0.932, LPIPS=0.147 Traditional 2DGS PNSR=31.68, SSIM= 0.933, LPIPS= 0.133 Ours, N = 4 PNSR=27.90, SSIM=0.883, LPIPS=0.198 Traditional 2DGS PNSR=28.15, SSIM= 0.888, LPIPS= 0.178 Ours...

    work page 2010

  3. [3]

    Mini-splatting: Represent- ing scenes with a constrained number of gaussians

    Guangchi Fang and Bing Wang. Mini-splatting: Represent- ing scenes with a constrained number of gaussians. arXiv [cs.CV], 2024. 2

    work page 2024

  4. [4]

    SRGS: Super- resolution 3D gaussian splatting

    Xiang Feng, Yongbo He, Yubo Wang, Yan Yang, Zhenzhong Kuang, Yu Jun, Jianping Fan, and Jiajun Ding. SRGS: Super- resolution 3D gaussian splatting. arXiv [cs.CV], 2024. 2

    work page 2024

  5. [5]

    Automatic generation and detection of highly reliable fiducial markers under occlusion

    S Garrido-Jurado, R Mu ˜noz-Salinas, F J Madrid-Cuevas, and M J Mar ´ın-Jim´enez. Automatic generation and detection of highly reliable fiducial markers under occlusion. Pattern Recognit., 47(6):2280–2292, 2014. 7

    work page 2014

  6. [6]

    2d gaussian splatting for geometrically accu- rate radiance fields

    Binbin Huang, Zehao Yu, Anpei Chen, Andreas Geiger, and Shenghua Gao. 2d gaussian splatting for geometrically accu- rate radiance fields. In ACM SIGGRAPH 2024 Conference Papers, New York, NY , USA, 2024. Association for Com- puting Machinery. 1, 3, 5

    work page 2024

  7. [7]

    3DGS-LM: Faster gaussian-splatting optimization with levenberg-marquardt

    Lukas H ¨ollein, Aljaˇz Boˇziˇc, Michael Zollh¨ofer, and Matthias Nießner. 3DGS-LM: Faster gaussian-splatting optimization with levenberg-marquardt. arXiv [cs.CV], 2024. 2

    work page 2024

  8. [8]

    GaussianShader: 3D gaussian splatting with shading functions for reflective surfaces

    Yingwenqi Jiang, Jiadong Tu, Yuan Liu, Xifeng Gao, Xiaox- iao Long, Wenping Wang, and Yuexin Ma. GaussianShader: 3D gaussian splatting with shading functions for reflective surfaces. arXiv [cs.CV], 2023. 5

    work page 2023

Show all 29 references
  1. [9]

    3D gaussian splatting for real-time radiance field rendering

    Bernhard Kerbl, Georgios Kopanas, Thomas Leimkuehler, and George Drettakis. 3D gaussian splatting for real-time radiance field rendering. ACM Trans. Graph., 42(4):1–14,

  2. [10]

    Adam: A method for stochastic optimization

    Diederik P Kingma and Jimmy Ba. Adam: A method for stochastic optimization. arXiv [cs.LG], 2014. 3

    2014

  3. [11]

    3D-HGS: 3D half-gaussian splatting

    Haolin Li, Jinyang Liu, Mario Sznaier, and Octavia Camps. 3D-HGS: 3D half-gaussian splatting. arXiv [cs.CV], 2024. 2

    2024

  4. [12]

    Real-time high-resolution background matting

    Shanchuan Lin, Andrey Ryabtsev, Soumyadip Sengupta, Brian L Curless, Steven M Seitz, and Ira Kemelmacher- Shlizerman. Real-time high-resolution background matting. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 8762–8771, 2021. 7

    2021

  5. [13]

    Mix- ture of volumetric primitives for efficient neural rendering

    Stephen Lombardi, Tomas Simon, Gabriel Schwartz, Michael Zollhoefer, Yaser Sheikh, and Jason Saragih. Mix- ture of volumetric primitives for efficient neural rendering. ACM Trans. Graph., 40(4):1–13, 2021. 1

    2021

  6. [14]

    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 synthe- sis. Commun. ACM, 65(1):99–106, 2021. 1, 2

    2021

  7. [15]

    StopThePop: Sorted gaussian splatting for view-consistent real-time rendering

    Lukas Radl, Michael Steiner, Mathias Parger, Alexan- der Weinrauch, Bernhard Kerbl, and Markus Steinberger. StopThePop: Sorted gaussian splatting for view-consistent real-time rendering. arXiv [cs.GR], 2024. 2

    2024

  8. [16]

    GStex: Per-primitive tex- turing of 2D gaussian splatting for decoupled appearance and geometry modeling

    Victor Rong, Jingxiang Chen, Sherwin Bahmani, Kiriakos N Kutulakos, and David B Lindell. GStex: Per-primitive tex- turing of 2D gaussian splatting for decoupled appearance and geometry modeling. arXiv [cs.CV], 2024. 2

    2024

  9. [17]

    Gaussian splatting on the move: Blur and rolling shutter compensation for natural camera motion

    Otto Seiskari, Jerry Ylilammi, Valtteri Kaatrasalo, Pekka Rantalankila, Matias Turkulainen, Juho Kannala, Esa Rahtu, and Arno Solin. Gaussian splatting on the move: Blur and rolling shutter compensation for natural camera motion. arXiv [cs.CV], 2024. 2

    2024

  10. [18]

    Image quality assessment: from error vis- ibility to structural similarity

    Zhou Wang, Alan Conrad Bovik, Hamid Rahim Sheikh, and Eero P Simoncelli. Image quality assessment: from error vis- ibility to structural similarity. IEEE Trans. Image Process., 13(4):600–612, 2004. 5

    2004

  11. [19]

    Texture-GS: Disentangling the geometry and texture for 3D gaussian splatting editing

    Tian-Xing Xu, Wenbo Hu, Yu-Kun Lai, Ying Shan, and Song-Hai Zhang. Texture-GS: Disentangling the geometry and texture for 3D gaussian splatting editing. arXiv [cs.CV],

  12. [20]

    Multi-scale 3D gaussian splatting for anti-aliased rendering

    Zhiwen Yan, Weng Fei Low, Yu Chen, and Gim Hee Lee. Multi-scale 3D gaussian splatting for anti-aliased rendering. arXiv [cs.CV], 2023. 2

    2023

  13. [21]

    gsplat: An open-source library for Gaussian splatting

    Vickie Ye, Ruilong Li, Justin Kerr, Matias Turkulainen, Brent Yi, Zhuoyang Pan, Otto Seiskari, Jianbo Ye, Jeffrey Hu, Matthew Tancik, and Angjoo Kanazawa. gsplat: An open-source library for Gaussian splatting. arXiv preprint arXiv:2409.06765, 2024. 3, 1

    2024 arXiv

  14. [22]

    AbsGS: Recovering fine details for 3D gaussian splat- ting

    Zongxin Ye, Wenyu Li, Sidun Liu, Peng Qiao, and Yong Dou. AbsGS: Recovering fine details for 3D gaussian splat- ting. arXiv [cs.CV], 2024. 2

    2024

  15. [23]

    2DGH: 2D gaussian-hermite splatting for high-quality ren- dering and better geometry reconstruction

    Ruihan Yu, Tianyu Huang, Jingwang Ling, and Feng Xu. 2DGH: 2D gaussian-hermite splatting for high-quality ren- dering and better geometry reconstruction. arXiv [cs.CV],

  16. [24]

    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. arXiv [cs.CV], 2023. 2

    2023

  17. [25]

    FreGS: 3D gaussian splatting with progressive frequency regularization

    Jiahui Zhang, Fangneng Zhan, Muyu Xu, Shijian Lu, and Eric Xing. FreGS: 3D gaussian splatting with progressive frequency regularization. arXiv [cs.CV], 2024. 2

    2024

  18. [26]

    The unreasonable effectiveness of deep features as a perceptual metric

    Richard Zhang, Phillip Isola, Alexei A Efros, Eli Shechtman, and Oliver Wang. The unreasonable effectiveness of deep features as a perceptual metric. CVPR, 2018. 5

    2018

  19. [27]

    EW A vol- ume splatting

    M Zwicker, H Pfister, J van Baar, and M Gross. EW A vol- ume splatting. In Proceedings Visualization, 2001. VIS ’01., pages 29–538. IEEE, 2001. 2

    2001

  20. [28]

    EW A splat- ting

    M Zwicker, H Pfister, J van Baar, and M Gross. EW A splat- ting. IEEE Trans. Vis. Comput. Graph., 8(3):223–238, 2002. 2 Gaussian Billboards: Expressive 2D Gaussian Splatting with Textures Supplementary Material

    2002

  21. [29]

    3.2 we introduce the basic algorithm of Gaussian Billboards, enhancing 2DGS with spatially-varying tex- tures

    Algorithm Details In Sec. 3.2 we introduce the basic algorithm of Gaussian Billboards, enhancing 2DGS with spatially-varying tex- tures. Here we give additional implementation details on how to integrate this into the gsplat kernel [21]. In Algorithm 1 we show where the biline...

Pith tools

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