pith. sign in

arxiv: 2606.32036 · v1 · pith:UKCJF6D2new · submitted 2026-06-30 · 💻 cs.CV

PointSplat: Compact Gaussian Splatting via Human-Centric Prediction

Yujie Guo , Yudong Jin , Lingteng Qiu , Zehong Shen , Zhen Xu , Jing Zhang , Xianchao Shen , Hujun Bao
show 2 more authors
Sida Peng Xiaowei Zhou
This is my paper

Pith reviewed 2026-07-01 05:23 UTC · model grok-4.3

classification 💻 cs.CV
keywords Gaussian splatting3D human reconstructionpoint cloudneural renderingcompact representationfeed-forward reconstructionray castingnovel view synthesis
0
0 comments X

The pith

PointSplat infers compact Gaussian primitives directly from 3D point sets for human rendering.

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

The paper establishes that shifting from view-centric to human-centric prediction in 3D space reduces inter-view redundancy in Gaussian splatting. It does this by starting from an input point cloud, estimating a coarse proxy, ray-casting to prune points and link 2D-3D correspondences, then running a Point-Image Transformer to output Gaussian attributes in one pass. The result is fewer total Gaussians confined to foreground regions, which the authors show improves novel-view quality. This matters for live streaming and similar settings where transmission bandwidth and compute are constrained. Experiments across datasets indicate the approach stays effective when the number of input views or their resolution changes.

Core claim

PointSplat directly infers Gaussian primitives from an input point set by first estimating a coarse geometric proxy and performing ray casting to prune redundant points and establish explicit 2D-3D correspondences, then employing a Point-Image Transformer to fuse appearance and geometry features and predict Gaussian attributes in a single forward pass, thereby restricting predictions to foreground regions of interest and yielding a more compact representation.

What carries the argument

Point-Image Transformer that fuses appearance and geometry features after ray-casting pruning on a coarse geometric proxy to predict Gaussian attributes from point sets.

If this is right

  • The total number of Gaussians needed for high-fidelity human rendering drops because predictions are limited to foreground regions.
  • Novel-view rendering quality rises relative to methods that encode the same content repeatedly across views.
  • Performance holds across varying input view counts and image resolutions without retraining.
  • Lower data volume supports real-time transmission in bandwidth-limited immersive systems.

Where Pith is reading between the lines

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

  • The same pruning-plus-transformer pipeline could be tested on non-human scenes if foreground identification generalizes.
  • Integration with existing point-cloud capture pipelines would allow end-to-end 3D content pipelines without intermediate multi-view encoding.
  • Reduced Gaussian count per subject could lower memory and decode costs in downstream AR or VR viewers.

Load-bearing premise

Estimating a coarse geometric proxy followed by ray casting can reliably prune redundant points and establish 2D-3D correspondences without losing critical appearance or geometry details needed for accurate Gaussian attribute prediction.

What would settle it

Rendering quality that falls below view-centric baselines on the same input views and resolutions when the number of views is reduced would show the claimed robustness does not hold.

Figures

Figures reproduced from arXiv: 2606.32036 by Hujun Bao, Jing Zhang, Lingteng Qiu, Sida Peng, Xianchao Shen, Xiaowei Zhou, Yudong Jin, Yujie Guo, Zehong Shen, Zhen Xu.

Figure 1
Figure 1. Figure 1: Our method synthesizes high-fidelity novel views from sparse views by predict￾ing a compact 3D Gaussian [17] representation. The bottom row shows representative results of our method on diverse human performances. employ dense camera arrays for view interpolation, can achieve high-quality rendering with low latency. However, their reliance on costly capture hardware severely limits scalability and deployme… view at source ↗
Figure 2
Figure 2. Figure 2: Comparison between view-centric and human-centric prediction. [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: It begins with the extraction of appearance features from 2D images [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figure 3
Figure 3. Figure 3: Pipeline Overview. Given calibrated multi-view images and masks, PointSplat reconstructs a compact 3D Gaussian (3DGS) representation in a feed￾forward step. We first extract appearance features from input images together with Plücker ray embeddings. A coarse geometric proxy is then estimated via space carv￾ing, followed by voxelization for structured representation. For each anchor view, we perform ray cas… view at source ↗
Figure 4
Figure 4. Figure 4: Qualitative comparison across datasets. Our method produces compact and complete 3D reconstructions that yield clean and consistent novel-view renderings. View-centric baselines (e.g., GPS-Gaussian, DepthSplat) show redundant or missing ge￾ometry, resulting in floating artifacts, while LVSM exhibits over-smoothed and mosaic￾like appearances. it delivers similar visual fidelity with a fraction of the comput… view at source ↗
Figure 5
Figure 5. Figure 5: Qualitative examples of novel view synthesis on the DyMVHumans [PITH_FULL_IMAGE:figures/full_fig_p012_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Ablation study of modules. Removing our proposed components leads to visible artifacts in the rendered results, especially in challenging regions such as the face [PITH_FULL_IMAGE:figures/full_fig_p013_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: Qualitative comparison of dense input views [PITH_FULL_IMAGE:figures/full_fig_p021_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: Results on synthetic datasets. Additional qualitative examples on synthetic scenes. Our method produces sharp renderings with stable geometry and consistent textures under viewpoint changes. Random Uniform FVS GT [PITH_FULL_IMAGE:figures/full_fig_p022_8.png] view at source ↗
Figure 9
Figure 9. Figure 9: Results under different camera input setups. [PITH_FULL_IMAGE:figures/full_fig_p022_9.png] view at source ↗
Figure 10
Figure 10. Figure 10: Mask robustness under random foreground-pixel dropping. [PITH_FULL_IMAGE:figures/full_fig_p023_10.png] view at source ↗
Figure 11
Figure 11. Figure 11: Comparison of SMPL and visual-hull initialization. [PITH_FULL_IMAGE:figures/full_fig_p024_11.png] view at source ↗
Figure 12
Figure 12. Figure 12: Additional analysis of point-set prediction. [PITH_FULL_IMAGE:figures/full_fig_p024_12.png] view at source ↗
read the original abstract

Producing 3D human representations from input views on the fly is essential for immersive live streaming systems, where representation compactness is as critical as high fidelity given limited computational power and transmission bandwidth. Although recent feed-forward reconstruction methods achieve impressive quality through the view-centric prediction of 3D representations, they repeatedly encode the same subject content across multiple views, leading to significant inter-view redundancy. Our key insight is to perform predictions directly in 3D space, enabling the network to learn and produce a highly compact representation. To this end, we propose PointSplat, a novel human-centric approach that directly infers Gaussian primitives from an input point set. The proposed method first estimates a coarse geometric proxy and performs ray casting to prune redundant points and establish explicit 2D--3D correspondences. Subsequently, it employs a Point-Image Transformer to fuse appearance and geometry features, predicting Gaussian attributes in a single forward pass. This design restricts predictions to foreground regions of interest, substantially reducing the total number of Gaussians while improving novel-view rendering quality. Extensive experiments demonstrate that PointSplat achieves higher efficiency and quality while exhibiting strong robustness to variations in view count and image resolution across multiple datasets.

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

1 major / 0 minor

Summary. The manuscript introduces PointSplat, a feed-forward human-centric method for compact 3D Gaussian splatting. It directly predicts Gaussian primitives from an input point set by first estimating a coarse geometric proxy, applying ray casting to prune redundant points and establish 2D-3D correspondences, then using a Point-Image Transformer to fuse appearance and geometry features and predict all Gaussian attributes in one forward pass. The design is claimed to restrict predictions to foreground regions, yielding fewer Gaussians, higher novel-view rendering quality, and robustness to changes in view count and image resolution across datasets.

Significance. If the efficiency and quality claims are substantiated, the work could meaningfully advance real-time 3D human reconstruction for bandwidth-constrained applications such as live streaming by eliminating inter-view redundancy through explicit 3D-space prediction rather than repeated view-centric encoding.

major comments (1)
  1. [Method (pruning and correspondence stage)] The ray-casting pruning step after coarse proxy estimation is load-bearing for the dual claim of fewer Gaussians and improved quality, yet the manuscript provides no quantitative validation (e.g., fraction of points retained, PSNR change when the pruning module is disabled, or ablation across view counts) that critical geometry or appearance details are preserved; without such checks the central compactness-plus-quality result remains unverified.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for the constructive feedback on our manuscript. We address the major comment below and will revise the paper accordingly.

read point-by-point responses

  1. Referee: [Method (pruning and correspondence stage)] The ray-casting pruning step after coarse proxy estimation is load-bearing for the dual claim of fewer Gaussians and improved quality, yet the manuscript provides no quantitative validation (e.g., fraction of points retained, PSNR change when the pruning module is disabled, or ablation across view counts) that critical geometry or appearance details are preserved; without such checks the central compactness-plus-quality result remains unverified.

    Authors: We agree that the manuscript would be strengthened by explicit quantitative validation of the ray-casting pruning step. In the revised version we will add an ablation that reports (i) the fraction of points retained after pruning, (ii) the PSNR change when the pruning module is disabled, and (iii) performance across varying input view counts. These results will confirm that critical geometry and appearance details are preserved while the pruning contributes to the reported compactness and quality gains. revision: yes

Circularity Check

0 steps flagged

No significant circularity; architectural pipeline is self-contained

full rationale

The paper presents PointSplat as a feed-forward architecture that estimates a coarse geometric proxy, applies ray casting for pruning and correspondence, then uses a Point-Image Transformer to predict Gaussian attributes directly from the input point set. No equations, fitted parameters renamed as predictions, or self-citation chains are shown that would reduce the claimed compactness or quality gains to the inputs by construction. The pruning step is described as an explicit design choice to restrict predictions to foreground regions rather than a derived necessity that loops back on itself. The derivation chain remains independent of the target results and does not invoke uniqueness theorems or ansatzes from prior author work.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Abstract alone supplies no explicit free parameters, axioms, or invented entities; the method description implies standard neural-network components without detailing any ad-hoc constants or new postulated objects.

pith-pipeline@v0.9.1-grok · 5771 in / 1138 out tokens · 29296 ms · 2026-07-01T05:23:03.795041+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

57 extracted references · 25 canonical work pages · 5 internal anchors

  1. [1]

    Ba, J.L., Kiros, J.R., Hinton, G.E.: Layer normalization (2016), arXiv:1607.06450

    work page internal anchor Pith review Pith/arXiv arXiv 2016

  2. [2]

    In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) (2024)

    Charatan, D., Li, S., Tagliasacchi, A., Sitzmann, V.: pixelsplat: 3d gaussian splats from image pairs for scalable generalizable 3d reconstruction. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) (2024)

    2024

  3. [3]

    In: Proceedings of the The International Conference on Learning Representations (ICLR) (2025)

    Chen, Y., Wu, Q., Li, M., Lin, W., Harandi, M., Cai, J.: Fast feedforward 3d gaus- sian splatting compression. In: Proceedings of the The International Conference on Learning Representations (ICLR) (2025)

    2025

  4. [4]

    arXiv preprint arXiv:2403.14627 (2024)

    Chen, Y., Xu, H., Zheng, C., Zhuang, B., Pollefeys, M., Geiger, A., Cham, T.J., Cai, J.: Mvsplat: Efficient 3d gaussian splatting from sparse multi-view images. arXiv preprint arXiv:2403.14627 (2024)

    work page arXiv 2024

  5. [5]

    arXiv preprintarXiv:2307.10173(2023)

    Cheng, W., Chen, R., Yin, W., Fan, S., Chen, K., He, H., Luo, H., Cai, Z., Wang, J., Gao, Y., Yu, Z., Lin, Z., Ren, D., Yang, L., Liu, Z., Loy, C.C., Qian, C., Wu, W., Lin, D., Dai, B., Lin, K.Y.: Dna-rendering: A diverse neural actor repository for high-fidelity human-centric rendering. arXiv preprintarXiv:2307.10173(2023)

    work page arXiv 2023

  6. [6]

    In: Proceedings of the The International Conference on Learning Repre- sentations (ICLR) (2024)

    Dao, T.: FlashAttention-2: Faster attention with better parallelism and work par- titioning. In: Proceedings of the The International Conference on Learning Repre- sentations (ICLR) (2024)

    2024

  7. [7]

    In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) (2022)

    Fridovich-Keil, S., Yu, A., Tancik, M., Chen, Q., Recht, B., Kanazawa, A.: Plenox- els: Radiance fields without neural networks. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) (2022)

    2022

  8. [8]

    Proceedings of the NeurIPS (NeurIPS) (2024)

    Gao, R., Holynski, A., Henzler, P., Brussee, A., Martin-Brualla, R., Srinivasan, P.P., Barron, J.T., Poole, B.: Cat3d: Create anything in 3d with multi-view diffu- sion models. Proceedings of the NeurIPS (NeurIPS) (2024)

    2024

  9. [9]

    In: Pro- ceedings of the Annual Conference on Computer Graphics and Interactive Tech- niques (SIGGRAPH)

    Gortler, S.J., Grzeszczuk, R., Szeliski, R., Cohen, M.F.: The lumigraph. In: Pro- ceedings of the Annual Conference on Computer Graphics and Interactive Tech- niques (SIGGRAPH). pp. 43–54. SIGGRAPH ’96 (1996).https://doi.org/10. 1145/237170.237200

    work page arXiv 1996

  10. [10]

    Henry, A., Dachapally, P.R., Pawar, S., Chen, Y.: Query-key normalization for transformers (2020), arXiv:2010.04245

    work page arXiv 2020

  11. [11]

    LRM: Large Reconstruction Model for Single Image to 3D

    Hong, Y., Zhang, K., Gu, J., Bi, S., Zhou, Y., Liu, D., Liu, F., Sunkavalli, K., Bui, T., Tan, H.: Lrm: Large reconstruction model for single image to 3d. arXiv preprint arXiv:2311.04400 (2023)

    work page internal anchor Pith review Pith/arXiv arXiv 2023

  12. [12]

    arXiv preprint (2023)

    Hu, S., Liu, Z.: Gauhuman: Articulated gaussian splatting for real-time 3d human rendering. arXiv preprint (2023)

    2023

  13. [13]

    ACM Transactions on Graphics (TOG)42(4), 1–12 (2023).https://doi.org/10

    Işık, M., Rünz, M., Georgopoulos, M., Khakhulin, T., Starck, J., Agapito, L., Nießner, M.: Humanrf: High-fidelity neural radiance fields for humans in motion. ACM Transactions on Graphics (TOG)42(4), 1–12 (2023).https://doi.org/10. 1145/3592415 16 Y. Guo et al

    2023

  14. [14]

    arXiv preprint arXiv:2505.23716 (2025)

    Jiang, L., Mao, Y., Xu, L., Lu, T., Ren, K., Jin, Y., Xu, X., Yu, M., Pang, J., Zhao, F., et al.: Anysplat: Feed-forward 3d gaussian splatting from unconstrained views. arXiv preprint arXiv:2505.23716 (2025)

    work page arXiv 2025

  15. [15]

    In: Pro- ceedings of the The International Conference on Learning Representations (ICLR) (2025)

    Jin, H., Jiang, H., Tan, H., Zhang, K., Bi, S., Zhang, T., Luan, F., Snavely, N., Xu, Z.: Lvsm: A large view synthesis model with minimal 3d inductive bias. In: Pro- ceedings of the The International Conference on Learning Representations (ICLR) (2025)

    2025

  16. [16]

    In: Proceedings of the IEEE International Con- ference on Computer Vision (ICCV) (2025)

    Jin, Y., Peng, S., Wang, X., Xie, T., Xu, Z., Yang, Y., Shen, Y., Bao, H., Zhou, X.: Diffuman4d: 4d consistent human view synthesis from sparse-view videos with spatio-temporal diffusion models. In: Proceedings of the IEEE International Con- ference on Computer Vision (ICCV) (2025)

    2025

  17. [17]

    ACM Transactions on Graphics (TOG)42(4) (2023)

    Kerbl, B., Kopanas, G., Leimkühler, T., Drettakis, G.: 3d gaussian splatting for real-time radiance field rendering. ACM Transactions on Graphics (TOG)42(4) (2023)

    2023

  18. [18]

    In: Proceedings of the IEEE International Conference on Computer Vision (ICCV)

    Kutulakos, K., Seitz, S.: A theory of shape by space carving. In: Proceedings of the IEEE International Conference on Computer Vision (ICCV). vol. 1, pp. 307–314 vol.1 (1999).https://doi.org/10.1109/ICCV.1999.791235

    work page doi:10.1109/iccv.1999.791235 1999

  19. [19]

    APSIPA Transactions on Signal and Information Processing4, e15 (2015).https://doi.org/10.1017/ATSIP.2015.18

    Lee, C.C., Tabatabai, A., Tashiro, K.: Free viewpoint video (fvv) survey and future research direction. APSIPA Transactions on Signal and Information Processing4, e15 (2015).https://doi.org/10.1017/ATSIP.2015.18

    work page doi:10.1017/atsip.2015.18 2015

  20. [20]

    In: Proceedings of the Annual Conference on Computer Graphics and Interactive Techniques (SIGGRAPH)

    Levoy, M., Hanrahan, P.: Light field rendering. In: Proceedings of the Annual Conference on Computer Graphics and Interactive Techniques (SIGGRAPH). pp. 31–42. SIGGRAPH ’96 (1996).https://doi.org/10.1145/237170.237199

    work page doi:10.1145/237170.237199 1996

  21. [21]

    arXiv preprint arXiv:2311.06214 (2023)

    Li, J., Tan, H., Zhang, K., Xu, Z., Luan, F., Xu, Y., Hong, Y., Sunkavalli, K., Shakhnarovich, G., Bi, S.: Instant3d: Fast text-to-3d with sparse-view generation and large reconstruction model. arXiv preprint arXiv:2311.06214 (2023)

    work page arXiv 2023

  22. [22]

    Depth Anything 3: Recovering the Visual Space from Any Views

    Lin, H., Chen, S., Liew, J.H., Chen, D.Y., Li, Z., Shi, G., Feng, J., Kang, B.: Depth anything 3: recovering the visual space from any views. arXiv preprint arXiv:2511.10647 (2025)

    work page internal anchor Pith review Pith/arXiv arXiv 2025

  23. [23]

    In: Proceedings of the ACM SIGGRAPH Conference and Exhibition on Computer Graphics and Interactive Techniques in Asia (SIGGRAPH Asia) (2022)

    Lin, H., Peng, S., Xu, Z., Yan, Y., Shuai, Q., Bao, H., Zhou, X.: Efficient neural radiance fields for interactive free-viewpoint video. In: Proceedings of the ACM SIGGRAPH Conference and Exhibition on Computer Graphics and Interactive Techniques in Asia (SIGGRAPH Asia) (2022)

    2022

  24. [24]

    In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)

    Lu, T., Yu, M., Xu, L., Xiangli, Y., Wang, L., Lin, D., Dai, B.: Scaffold-gs: Struc- tured 3d gaussians for view-adaptive rendering. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). pp. 20654– 20664 (2024)

    2024

  25. [25]

    ACM Transactions on Graphics (TOG)38(4), 1–14 (2019)

    Mildenhall, B., Srinivasan, P.P., Ortiz-Cayon, R., Kalantari, N.K., Ramamoorthi, R., Ng, R., Kar, A.: Local light field fusion: Practical view synthesis with pre- scriptive sampling guidelines. ACM Transactions on Graphics (TOG)38(4), 1–14 (2019)

    2019

  26. [26]

    Mildenhall, B., Srinivasan, P.P., Tancik, M., Barron, J.T., Ramamoorthi, R., Ng, R.: Nerf: representing scenes as neural radiance fields for view synthesis. Commun. ACM65(1), 99–106 (2021).https://doi.org/10.1145/3503250

    work page doi:10.1145/3503250 2021

  27. [27]

    ACM Transactions on Graphics (TOG)41(4) (2022).https://doi.org/10.1145/3528223.3530127

    Müller,T.,Evans,A.,Schied,C.,Keller,A.:Instantneuralgraphicsprimitiveswith a multiresolution hash encoding. ACM Transactions on Graphics (TOG)41(4) (2022).https://doi.org/10.1145/3528223.3530127

    work page doi:10.1145/3528223.3530127 2022

  28. [28]

    Ng, R., Levoy, M., Brédif, M., Duval, G., Horowitz, M., Hanrahan, P.: Light field photography with a hand-held plenoptic camera. Ph.D. thesis, Stanford university (2005) PointSplat: Compact Gaussian Splatting via Human-Centric Prediction 17

    2005

  29. [29]

    In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) (2021)

    Peng, S., Zhang, Y., Xu, Y., Wang, Q., Shuai, Q., Bao, H., Zhou, X.: Neural body: Implicit neural representations with structured latent codes for novel view synthesis of dynamic humans. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) (2021)

    2021

  30. [31]

    arXiv preprint arXiv:2503.10625 (2025)

    Qiu, L., Li, P., Li, H., Zuo, Q., Gu, X., Dong, Y., Yuan, W., Peng, R., Zhu, S., Han, X., Chen, G., Dong, Z.: Lhm++: An efficient large human reconstruction model for pose-free images to 3d. arXiv preprint arXiv:2503.10625 (2025)

    work page arXiv 2025

  31. [32]

    arXiv preprint arXiv:2403.17898 (2024)

    Ren, K., Jiang, L., Lu, T., Yu, M., Xu, L., Ni, Z., Dai, B.: Octree-gs: Towards consistent real-time rendering with lod-structured 3d gaussians. arXiv preprint arXiv:2403.17898 (2024)

    work page arXiv 2024

  32. [33]

    Renderpeople: Renderpeople.https://renderpeople.com/3d-people(2018)

    2018

  33. [34]

    Szymanowicz, S., Zhang, J.Y., Srinivasan, P., Gao, R., Brussee, A., Holynski, A., Martin-Brualla, R., Barron, J.T., Henzler, P.: Bolt3D: Generating 3D Scenes in Seconds.In:ProceedingsoftheIEEEInternationalConferenceonComputerVision (ICCV) (2025)

    2025

  34. [35]

    arXiv preprint arXiv:2402.05054 (2024)

    Tang, J., Chen, Z., Chen, X., Wang, T., Zeng, G., Liu, Z.: Lgm: Large multi- view gaussian model for high-resolution 3d content creation. arXiv preprint arXiv:2402.05054 (2024)

    work page arXiv 2024

  35. [36]

    Team, T.H.: Hunyuan3d 2.0: Scaling diffusion models for high resolution textured 3d assets generation (2025), arXiv:2501.12202

    work page internal anchor Pith review Pith/arXiv arXiv 2025

  36. [37]

    In: Proceedings of the IEEE/CVF Confer- ence on Computer Vision and Pattern Recognition (CVPR) (2025)

    Wang, J., Chen, M., Karaev, N., Vedaldi, A., Rupprecht, C., Novotny, D.: Vggt: Visual geometry grounded transformer. In: Proceedings of the IEEE/CVF Confer- ence on Computer Vision and Pattern Recognition (CVPR) (2025)

    2025

  37. [38]

    In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) (2021)

    Wang, Q., Wang, Z., Genova, K., Srinivasan, P., Zhou, H., Barron, J.T., Martin- Brualla, R., Snavely, N., Funkhouser, T.: Ibrnet: Learning multi-view image-based rendering. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) (2021)

    2021

  38. [39]

    Wang, S., Leroy, V., Cabon, Y., Chidlovskii, B., Revaud, J.: Dust3r: Geometric 3d vision made easy (2023), arXiv:2312.14132

    work page arXiv 2023

  39. [40]

    Wang, Y., Zhou, J., Zhu, H., Chang, W., Zhou, Y., Li, Z., Chen, J., Pang, J., Shen, C., He, T.:π3: Scalable permutation-equivariant visual geometry learning (2025), arXiv:2507.13347

    work page internal anchor Pith review Pith/arXiv arXiv 2025

  40. [41]

    C.; Sheikh, H

    Wang, Z., Bovik, A., Sheikh, H., Simoncelli, E.: Image quality assessment: from error visibility to structural similarity. IEEE Transactions on Image Processing (TIP)13(4), 600–612 (2004).https://doi.org/10.1109/TIP.2003.819861

    work page doi:10.1109/tip.2003.819861 2004

  41. [42]

    arXiv preprint arXiv:2405.14832 (2024)

    Wu, S., Lin, Y., Zhang, F., Zeng, Y., Xu, J., Torr, P., Cao, X., Yao, Y.: Direct3d: Scalable image-to-3d generation via 3d latent diffusion transformer. arXiv preprint arXiv:2405.14832 (2024)

    work page arXiv 2024

  42. [43]

    In: Proceed- ings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) (2025)

    Xiao, J., Zhang, Q., Nie, Y., Zhu, L., Zheng, W.S.: RoGSplat: Learning robust gen- eralizable human gaussian splatting from sparse multi-view images. In: Proceed- ings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) (2025)

    2025

  43. [44]

    In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) (2025)

    Xu, H., Peng, S., Wang, F., Blum, H., Barath, D., Geiger, A., Pollefeys, M.: Depth- splat: Connecting gaussian splatting and depth. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) (2025)

    2025

  44. [45]

    In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)

    Xu, Q., Xu, Z., Philip, J., Bi, S., Shu, Z., Sunkavalli, K., Neumann, U.: Point-nerf: Point-based neural radiance fields. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). pp. 5438–5448 (2022) 18 Y. Guo et al

    2022

  45. [46]

    ACM Transactions on Graphics (TOG)43(6) (2024)

    Xu, Z., Xu, Y., Yu, Z., Peng, S., Sun, J., Bao, H., Zhou, X.: Representing long volumetric video with temporal gaussian hierarchy. ACM Transactions on Graphics (TOG)43(6) (2024)

    2024

  46. [47]

    In: Proceedings of the The International Conference on Learning Representations (ICLR) (2024)

    Yang, Z., Yang, H., Pan, Z., Zhang, L.: Real-time photorealistic dynamic scene representation and rendering with 4d gaussian splatting. In: Proceedings of the The International Conference on Learning Representations (ICLR) (2024)

    2024

  47. [48]

    In: Proceedings of the European Conference on Computer Vision (ECCV) (2018)

    Yao, Y., Luo, Z., Li, S., Fang, T., Quan, L.: Mvsnet: Depth inference for unstruc- tured multi-view stereo. In: Proceedings of the European Conference on Computer Vision (ECCV) (2018)

    2018

  48. [49]

    Yinghao, X., Zifan, S., Wang, Y., Hansheng, C., Ceyuan, Y., Sida, P., Yujun, S., Gordon, W.: Grm: Large gaussian reconstruction model for efficient 3d reconstruc- tion and generation (2024), arXiv:2403.14621

    work page arXiv 2024

  49. [50]

    In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) (2021)

    Yu, T., Zheng, Z., Guo, K., Liu, P., Dai, Q., Liu, Y.: Function4d: Real-time human volumetric capture from very sparse consumer rgbd sensors. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) (2021)

    2021

  50. [51]

    ACM Transactions on Graphics (TOG)42(4) (2023).https://doi.org/10.1145/3592442

    Zhang, B., Tang, J., Nießner, M., Wonka, P.: 3dshape2vecset: A 3d shape repre- sentation for neural fields and generative diffusion models. ACM Transactions on Graphics (TOG)42(4) (2023).https://doi.org/10.1145/3592442

    work page doi:10.1145/3592442 2023

  51. [52]

    In: Proceedings of the European Conference on Computer Vision (ECCV) (2024)

    Zhang, K., Bi, S., Tan, H., Xiangli, Y., Zhao, N., Sunkavalli, K., Xu, Z.: Gs- lrm: Large reconstruction model for 3d gaussian splatting. In: Proceedings of the European Conference on Computer Vision (ECCV) (2024)

    2024

  52. [53]

    ACM Transactions on Graphics (TOG)43(4) (2024).https: //doi.org/10.1145/3658146

    Zhang, L., Wang, Z., Zhang, Q., Qiu, Q., Pang, A., Jiang, H., Yang, W., Xu, L., Yu, J.: Clay: A controllable large-scale generative model for creating high- quality 3d assets. ACM Transactions on Graphics (TOG)43(4) (2024).https: //doi.org/10.1145/3658146

    work page doi:10.1145/3658146 2024

  53. [54]

    In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) (2018)

    Zhang, R., Isola, P., Efros, A.A., Shechtman, E., Wang, O.: The unreasonable effec- tiveness of deep features as a perceptual metric. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) (2018)

    2018

  54. [55]

    Proceedings of the NeurIPS (NeurIPS) (2024)

    Zhang, S., Fei, X., Liu, F., Song, H., Duan, Y.: Gaussian graph network: Learn- ing efficient and generalizable gaussian representations from multi-view images. Proceedings of the NeurIPS (NeurIPS) (2024)

    2024

  55. [56]

    In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) (2024)

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

    2024

  56. [57]

    In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) (2024)

    Zheng, X., Liao, L., Li, X., Jiao, J., Wang, R., Gao, F., Wang, S., Wang, R.: Pku-dymvhumans: A multi-view video benchmark for high-fidelity dynamic human modeling. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) (2024)

    2024

  57. [58]

    Random” indicates ran- domly sampled views. “Uniform

    Ziwen, C., Tan, H., Zhang, K., Bi, S., Luan, F., Hong, Y., Fuxin, L., Xu, Z.: Long- lrm: Long-sequence large reconstruction model for wide-coverage gaussian splats. In: Proceedings of the IEEE International Conference on Computer Vision (ICCV) (2025) PointSplat: Compact Gaussian Splatting via Human-Centric Prediction 19 Supplementary Material A Method Det...

    2025