pith. sign in

arxiv: 2605.22342 · v1 · pith:DARLFUKLnew · submitted 2026-05-21 · 💻 cs.CV · cs.AI

4D-GSW: Kinematic-Aware Spatio-Temporal Consistent Watermarking for 4D Gaussian Splatting

Sifan Zhou , Hang Zhang , Yuhang Wang , Ming Li This is my paper

Pith reviewed 2026-05-22 06:40 UTC · model grok-4.3

classification 💻 cs.CV cs.AI
keywords 4D Gaussian SplattingWatermarkingSpatio-Temporal ConsistencyKinematic AwarenessDynamic ReconstructionCopyright ProtectionMotion Coherence
0
0 comments X

The pith

A kinematic-aware method embeds watermarks in 4D Gaussian Splatting by gating at curvature-identified motion instants and synchronizing phases to avoid non-physical artifacts.

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

The paper aims to protect intellectual property in dynamic 4D reconstructions without causing visible flickering or loss of motion realism that plagues earlier hiding techniques. It identifies critical points in space-time motion using a curvature measure, then routes watermark changes only away from those points while aligning signals across nearby locations and times. A reader would care because 4D assets are increasingly used in animation, simulation, and virtual environments where both ownership and visual fidelity matter. If the approach holds, watermarks become harder to strip without damaging the underlying physical coherence of the scene.

Core claim

The central discovery is that watermark gradient injection can be adaptively gated at Spatio-Temporal Curvature-identified Dynamic Instants and synchronized through a joint HMM-MRF energy minimization, with anisotropic gradient routing keeping the embedding strictly decoupled from photometric reconstruction, thereby achieving robust copyright embedding that resists attacks while preserving high rendering quality and spatiotemporal consistency in 4D Gaussian Splatting.

What carries the argument

The Spatio-Temporal Curvature metric that locates Dynamic Instants for gated injection, paired with the HMM-MRF synchronization model and anisotropic gradient routing that maintains motion coherence and photometric decoupling.

If this is right

  • Watermarks survive removal attempts without triggering visible temporal inconsistencies in rendered output.
  • Motion manifolds stay physically plausible because injection is withheld precisely at high-curvature instants.
  • Global phase alignment across trajectories and neighborhoods produces consistent marks even under complex deformations.
  • Rendering quality metrics stay comparable to the original model because the embedding process is routed away from photometric gradients.

Where Pith is reading between the lines

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

  • Similar curvature-gated routing could be applied to other time-varying representations such as neural radiance fields for dynamic scenes.
  • The approach implies that future watermarking for physical simulations should treat kinematic structure as a first-class constraint rather than an afterthought.
  • If the synchronization holds under real capture noise, it opens the possibility of watermarking live 4D reconstructions from sensors without post-processing.

Load-bearing premise

That gating watermark changes at curvature-identified motion instants and synchronizing them via HMM-MRF will avoid new artifacts while leaving physical motion trajectories intact.

What would settle it

Render a sequence of watermarked 4D Gaussian Splatting frames into video and measure whether temporal flickering or FVD scores remain unchanged from the unwatermarked baseline after common attacks such as compression or noise addition.

Figures

Figures reproduced from arXiv: 2605.22342 by Hang Zhang, Ming Li, Sifan Zhou, Yuhang Wang.

Figure 1
Figure 1. Figure 1: Watermark framework comparison. (Upper) Conventional 3D-GS watermarking schemes necessitate independent fine-tuning for each individual frame in a dynamic sequence. (Bottom) Our 4D-GSW treats the 4D asset as a holistic kinematic manifold. By computing the Spatio￾Temporal Curvature (STC) of Gaussian trajectories, we adaptively gate the watermark gradients— prioritizing stable regions while shielding "Dynami… view at source ↗
Figure 2
Figure 2. Figure 2: 4D-GSW framework overview. Our pipeline extracts Spatio-Temporal Curvature (STC) from Gaussian trajectories to derive an embedding confidence weight. This weight guides a kinematic￾aware optimization that adaptively modulates watermark gradients, prioritizing stable radiance manifolds while shielding dynamic instants. The field is further regularized by a joint HMM-GMRF energy model and wavelet-domain supe… view at source ↗
Figure 3
Figure 3. Figure 3: Visualization of temporal consistency on the Guppie sequence. The baseline (b) suffers from severe temporal flickering and geometric jitter during rapid motion (red boxes). In contrast, our 4D-GSW (c) maintains high structural integrity and motion smoothness consistent with the baseline. 1) Spatio-Temporal and Cross-View Consistency: As shown in [PITH_FULL_IMAGE:figures/full_fig_p008_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Visualization across diverse dynamic assets. Comparison of models generated by SC4D, SC4D+Hidden, and our method across different viewpoints at the same time step. T1,V1 T1,V2 T1,V3 T2,V1 T2,V2 T2,V3 Input Image T1,V1 T1,V2 T1,V3 T2,V1 T2,V2 T2,V3 Input Image Ours SC4D SC4D+H Ours SC4D SC4D+H [PITH_FULL_IMAGE:figures/full_fig_p009_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Multi-view 4D Generation Results at Different Timestamps. Comparison of models generated by SC4D [37], SC4D+Hidden [48], and our method across different viewpoints at the different timestep. the Ground Truth. Furthermore, the 10× residual maps visually confirm our anisotropic gradient routing: the watermark signal is adaptively concentrated in stable, texture-rich regions and repelled from motion-sensitive… view at source ↗
Figure 6
Figure 6. Figure 6: Visualization across diverse dynamic assets. From left to right: (a) Ground Truth, (b) original SC4D, (c, d) SC4D+HiDDeN baseline and its 10× residual map, (e, f) our 4D-GSW and its 10× residual map. While the baseline introduces blurring at structural boundaries, 4D-GSW achieves near-lossless reconstruction. 17 [PITH_FULL_IMAGE:figures/full_fig_p017_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: Multi-view 4D generation at different timestamps. Qualitative comparison of SC4D, SC4D+Hidden, and our method under varying viewpoints at a fixed time step [PITH_FULL_IMAGE:figures/full_fig_p018_7.png] view at source ↗
read the original abstract

While 4D Gaussian Splatting (4DGS) has revolutionized high-fidelity dynamic reconstruction, safeguarding the intellectual property of these assets remains an open challenge. Conventional steganographic techniques often neglect the underlying kinematic manifolds, triggering non-physical artifacts such as severe temporal flickering and "FVD collapse". To address this, we propose \textbf{4D-GSW}, a kinematic-aware watermarking framework designed to embed robust copyright information while preserving high spatio-temporal consistency. Unlike prior 4D steganography that primarily focuses on opacity-guided invisibility, our approach explicitly addresses the physical coherence of motion trajectories. We introduce a \textbf{Spatio-Temporal Curvature (STC)} metric to identify "Dynamic Instants," adaptively gating watermark gradient injection to shield critical motion manifolds from non-physical perturbations. To ensure global coherence across complex deformations, we formulate a joint \textbf{HMM-MRF energy minimization} model that synchronizes watermark phases within both temporal trajectories and spatial neighborhoods. Furthermore, an \textbf{anisotropic gradient routing} mechanism ensures that watermark embedding remains strictly decoupled from photometric reconstruction fidelity. Extensive experiments have demonstrated the superior performance of our method in robustly hiding watermarks while resisting various attacks and maintaining high rendering quality and spatiotemporal consistency.

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

2 major / 2 minor

Summary. The manuscript proposes 4D-GSW, a kinematic-aware watermarking framework for 4D Gaussian Splatting. It introduces a Spatio-Temporal Curvature (STC) metric to identify 'Dynamic Instants' for adaptively gating watermark gradient injection, a joint HMM-MRF energy minimization model to synchronize watermark phases across temporal trajectories and spatial neighborhoods, and an anisotropic gradient routing mechanism claimed to keep watermark embedding strictly decoupled from photometric reconstruction fidelity. The central claim is that this approach embeds robust copyright information while preserving high spatio-temporal consistency, resisting attacks, and avoiding non-physical artifacts such as temporal flickering or FVD collapse, with extensive experiments demonstrating superior performance over prior 4D steganography methods.

Significance. If the central claims hold, the work would address a genuine gap in IP protection for dynamic 4D reconstructions by explicitly incorporating kinematic manifolds rather than relying solely on opacity-guided invisibility. The introduction of STC for motion-aware gating and HMM-MRF for cross-trajectory synchronization offers a targeted technical contribution to spatio-temporal consistency in Gaussian splatting watermarking.

major comments (2)
  1. [Abstract and Experiments] Abstract and Experiments: The abstract asserts 'superior performance' and 'resistance to various attacks' with 'high rendering quality and spatiotemporal consistency' but supplies no quantitative results, error bars, baseline comparisons, dataset details, or statistical tests. This absence is load-bearing for the superiority claim and prevents verification that the proposed STC gating, HMM-MRF synchronization, and anisotropic routing actually deliver the stated improvements.
  2. [Anisotropic gradient routing mechanism] Anisotropic gradient routing mechanism (described in abstract and method sections): The assertion that the mechanism 'ensures that watermark embedding remains strictly decoupled from photometric reconstruction fidelity' is not supported by an explicit orthogonality condition, such as a projection onto the tangent space of the motion manifold or a demonstrated null inner product between injected gradients and the STC vector field. Because 4DGS motion is represented via time-conditioned Gaussian parameters whose gradients are coupled through the volume rendering integral, the absence of this condition risks non-zero components along deformation trajectories, which would reintroduce the temporal flickering the method claims to eliminate.
minor comments (2)
  1. [Abstract] The acronym 'FVD' and the phrase 'FVD collapse' appear without definition or citation; they should be expanded and referenced on first use.
  2. [Method] Notation for the STC metric and the HMM-MRF energy terms should be introduced with explicit equations rather than descriptive prose to improve reproducibility.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive and detailed feedback. We address each major comment point by point below, indicating where revisions will be made to improve clarity and rigor.

read point-by-point responses

  1. Referee: [Abstract and Experiments] The abstract asserts 'superior performance' and 'resistance to various attacks' with 'high rendering quality and spatiotemporal consistency' but supplies no quantitative results, error bars, baseline comparisons, dataset details, or statistical tests. This absence is load-bearing for the superiority claim and prevents verification that the proposed STC gating, HMM-MRF synchronization, and anisotropic routing actually deliver the stated improvements.

    Authors: We agree that the abstract would be strengthened by including key quantitative highlights to support the claims of superiority. In the revised manuscript, we will update the abstract to incorporate specific metrics such as average PSNR improvements, watermark bit accuracy under representative attacks, and direct comparisons against prior 4D steganography baselines on the datasets used in our experiments. The experiments section already provides detailed tables with error bars, baseline results, and dataset specifications; we will add explicit cross-references from the abstract to these tables to make the evidence immediately verifiable. revision: yes

  2. Referee: [Anisotropic gradient routing mechanism] The assertion that the mechanism 'ensures that watermark embedding remains strictly decoupled from photometric reconstruction fidelity' is not supported by an explicit orthogonality condition, such as a projection onto the tangent space of the motion manifold or a demonstrated null inner product between injected gradients and the STC vector field. Because 4DGS motion is represented via time-conditioned Gaussian parameters whose gradients are coupled through the volume rendering integral, the absence of this condition risks non-zero components along deformation trajectories, which would reintroduce the temporal flickering the method claims to eliminate.

    Authors: We appreciate this observation on the need for a more explicit mathematical grounding. The anisotropic gradient routing mechanism employs the STC metric to identify and avoid directions aligned with critical motion manifolds, thereby routing watermark gradients preferentially along non-deforming components. To address the concern directly, we will add a formal derivation in the method section demonstrating that the routed gradient satisfies a null inner product with the STC vector field (i.e., an orthogonality condition with respect to the time-conditioned deformation parameters). This will clarify how the approach mitigates coupling through the volume rendering integral and prevents reintroduction of temporal artifacts. revision: yes

Circularity Check

0 steps flagged

No significant circularity; new metrics and mechanisms introduced without reduction to fitted inputs or self-citations

full rationale

The provided abstract and description introduce original components including the Spatio-Temporal Curvature (STC) metric for gating at Dynamic Instants, a joint HMM-MRF energy minimization model for synchronization, and an anisotropic gradient routing mechanism claimed to decouple watermark embedding from photometric fidelity. No equations, self-citations, or parameter-fitting steps are shown that would reduce any prediction or result to the inputs by construction. The central claims rest on experimental validation of robustness and consistency rather than mathematical equivalence to prior fitted values or author-specific priors. This matches the expectation that most papers lack circularity when they define new constructs independently.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Only the abstract is available; no concrete free parameters, axioms, or invented entities can be extracted or verified from the given text. The described components (STC metric, HMM-MRF model, anisotropic gradient routing) function as methodological inventions whose independence from prior literature cannot be checked.

pith-pipeline@v0.9.0 · 5767 in / 1242 out tokens · 38234 ms · 2026-05-22T06:40:30.638381+00:00 · methodology

discussion (0)

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

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

What do these tags mean?
matches
The paper's claim is directly supported by a theorem in the formal canon.
supports
The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends
The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses
The paper appears to rely on the theorem as machinery.
contradicts
The paper's claim conflicts with a theorem or certificate in the canon.
unclear
Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

48 extracted references · 48 canonical work pages

  1. [1]

    Per- gaussian embedding-based deformation for deformable 3d gaussian splatting

    Jeongmin Bae, Seoha Kim, Youngsik Yun, Hahyun Lee, Gun Bang, and Youngjung Uh. Per- gaussian embedding-based deformation for deformable 3d gaussian splatting. InEuropean Conference on Computer Vision (ECCV), 2024

    work page 2024

  2. [2]

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

    Jonathan T Barron, Ben Mildenhall, Matthew Tancik, Peter Hedman, Ricardo Martin-Brualla, and Pratul P Srinivasan. Mip-nerf: A multiscale representation for anti-aliasing neural radiance fields. InProceedings of the IEEE/CVF International Conference on Computer Vision, pages 5855–5864, 2021

    work page 2021

  3. [3]

    Zip-NeRF: Anti-aliased grid-based neural radiance fields

    Jonathan T Barron, Ben Mildenhall, Dor Verbin, Pratul P Srinivasan, and Peter Hedman. Zip-NeRF: Anti-aliased grid-based neural radiance fields. InProceedings of the IEEE/CVF International Conference on Computer Vision, pages 19697–19705, 2023

    work page 2023

  4. [4]

    Quo vadis, action recognition? a new model and the kinetics dataset

    Joao Carreira and Andrew Zisserman. Quo vadis, action recognition? a new model and the kinetics dataset. Inproceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pages 6299–6308, 2017

    work page 2017

  5. [5]

    HAIF-GS: Hierarchical and induced flow- guided gaussian splatting for dynamic scene

    Jianing Chen, Zehao Li, Yujun Cai, Hao Jiang, Chengxuan Qian, Juyuan Kang, Shuqin Gao, Honglong Zhao, Tianlu Mao, and Yucheng Zhang. HAIF-GS: Hierarchical and induced flow- guided gaussian splatting for dynamic scene. InThe Thirty-ninth Annual Conference on Neural Information Processing Systems, 2025

    work page 2025

  6. [6]

    Gaussianeditor: Swift and controllable 3d editing with gaussian splatting

    Yiwen Chen, Zilong Chen, Chi Zhang, Feng Wang, Xiaofeng Yang, Yikai Wang, Zhongang Cai, Lei Yang, Huaping Liu, and Guosheng Lin. Gaussianeditor: Swift and controllable 3d editing with gaussian splatting. InCVPR, 2024

    work page 2024

  7. [7]

    Guardsplat: Efficient and robust watermarking for 3d gaussian splatting

    Zixuan Chen, Guangcong Wang, Jiahao Zhu, Jianhuang Lai, and Xiaohua Xie. Guardsplat: Efficient and robust watermarking for 3d gaussian splatting. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 16325–16335, 2025

    work page 2025

  8. [8]

    Sinkhorn distances: Lightspeed computation of optimal transport.Advances in neural information processing systems, 26, 2013

    Marco Cuturi. Sinkhorn distances: Lightspeed computation of optimal transport.Advances in neural information processing systems, 26, 2013

    work page 2013

  9. [9]

    Splats in splats: Robust and effective 3d steganography towards gaussian splatting

    Yijia Guo, Wenkai Huang, Yang Li, Gaolei Li, Hang Zhang, Liwen Hu, Jianhua Li, Tiejun Huang, and Lei Ma. Splats in splats: Robust and effective 3d steganography towards gaussian splatting. InAAAI, 2026

    work page 2026

  10. [10]

    Vtgaussian-slam: Rgbd slam for large scale scenes with splatting view-tied 3d gaussians

    Pengchong Hu and Zhizhong Han. Vtgaussian-slam: Rgbd slam for large scale scenes with splatting view-tied 3d gaussians. InProceedings of the 42nd International Conference on Machine Learning, 2025

    work page 2025

  11. [11]

    Waterf: Robust watermarks in radiance fields for protection of copyrights

    Youngdong Jang, Seungwoo Bae, Jaekyung Kim, and Sangpil Yu. Waterf: Robust watermarks in radiance fields for protection of copyrights. InIEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2025

    work page 2025

  12. [12]

    3d-gsw: 3d gaussian splatting for robust watermarking

    Youngdong Jang, Hyunje Park, Feng Yang, Heeju Ko, Euijin Choo, and Sangpil Kim. 3d-gsw: 3d gaussian splatting for robust watermarking. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 5938–5948, 2025

    work page 2025

  13. [13]

    Consistent4d: Consistent 360° dynamic object generation from monocular video

    Yanqin Jiang, Li Zhang, Jin Gao, Weiming Hu, and Yao Yao. Consistent4d: Consistent 360° dynamic object generation from monocular video. InThe Twelfth International Conference on Learning Representations, 2024

    work page 2024

  14. [14]

    Hifi4g: High-fidelity human performance rendering via compact gaussian splatting

    Yuheng Jiang, Zhehao Shen, Penghao Wang, Zhuo Su, Yu Hong, Yingliang Zhang, Jingyi Yu, and Lan Xu. Hifi4g: High-fidelity human performance rendering via compact gaussian splatting. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 19734–19745, June 2024

    work page 2024

  15. [15]

    3d gaussian splatting for real-time radiance field rendering.ACM Transactions on Graphics, 2023

    Bernhard Kerbl, Georgios Kopanas, Thomas Leimkühler, and George Drettakis. 3d gaussian splatting for real-time radiance field rendering.ACM Transactions on Graphics, 2023. 11

    work page 2023

  16. [16]

    Feng, Zhiwen Fan, Panwang Pan, and Zhangyang Wang

    Chenxin Li, Brandon Y . Feng, Zhiwen Fan, Panwang Pan, and Zhangyang Wang. Steganerf: Embedding invisible information within neural radiance fields. InProceedings of the IEEE/CVF International Conference on Computer Vision (ICCV), 2023

    work page 2023

  17. [17]

    Gaussianmarker: Uncertainty-aware copyright protection of 3d gaussian splatting

    Hanzhi Li, Wenkang Li, Wenbo Hu, Kaidong Zhang, Yifan Zhao, Yuewen Liu, Yong Liu, and Boxin Wen. Gaussianmarker: Uncertainty-aware copyright protection of 3d gaussian splatting. InProceedings of the 32nd ACM International Conference on Multimedia, pages 2500–2509, 2024

    work page 2024

  18. [18]

    Hide-in-motion: Embedding steganographic copyright information into 4d gaussian splatting assets

    Hengyu Liu, Chenxin Li, Wentao Pan, Zhiqin Yang, Yifeng Yang, Yifan Liu, Wuyang Li, and Yixuan Yuan. Hide-in-motion: Embedding steganographic copyright information into 4d gaussian splatting assets. In2025 IEEE International Conference on Robotics and Automation (ICRA), pages 2694–2700, 2025

    work page 2025

  19. [19]

    Foundation model-guided gaussian splatting for 4d reconstruction of deformable tissues.IEEE Transactions on Medical Imaging, 2025

    Yifan Liu, Chenxin Li, Hengyu Liu, Chen Yang, and Yixuan Yuan. Foundation model-guided gaussian splatting for 4d reconstruction of deformable tissues.IEEE Transactions on Medical Imaging, 2025

    work page 2025

  20. [20]

    Dynamic 3d gaussians: Tracking by persistent dynamic view synthesis

    Jonathon Luiten, Georgios Kopanas, Bastian Leibe, and Deva Ramanan. Dynamic 3d gaussians: Tracking by persistent dynamic view synthesis. In3DV, 2024

    work page 2024

  21. [21]

    CopyRNeRF: Pro- tecting the copyright of neural radiance fields

    Ziyuan Luo, Qing Guo, Ka Chun Cheung, Simon See, and Renjie Wan. CopyRNeRF: Pro- tecting the copyright of neural radiance fields. InProceedings of the IEEE/CVF International Conference on Computer Vision, 2023

    work page 2023

  22. [22]

    NeRF: Representing scenes as neural radiance fields for view synthesis

    Ben Mildenhall, Pratul P Srinivasan, Matthew Tancik, Jonathan T Barron, Ravi Ramamoorthi, and Ren Ng. NeRF: Representing scenes as neural radiance fields for view synthesis. In European Conference on Computer Vision, pages 405–421. Springer, 2020

    work page 2020

  23. [23]

    Recent advances in optimal transport for machine learning.IEEE Transactions on Pattern Analysis and Machine Intelligence, 47(2):1161–1180, 2025

    Eduardo Fernandes Montesuma, Fred Maurice Ngolè Mboula, and Antoine Souloumiac. Recent advances in optimal transport for machine learning.IEEE Transactions on Pattern Analysis and Machine Intelligence, 47(2):1161–1180, 2025

    work page 2025

  24. [24]

    Barron, Sofien Bouaziz, Dan B Goldman, Steven M

    Keunhong Park, Utkarsh Sinha, Jonathan T. Barron, Sofien Bouaziz, Dan B Goldman, Steven M. Seitz, and Ricardo Martin-Brualla. Nerfies: Deformable neural radiance fields.ICCV, 2021

    work page 2021

  25. [25]

    D-NeRF: Neural Radiance Fields for Dynamic Scenes

    Albert Pumarola, Enric Corona, Gerard Pons-Moll, and Francesc Moreno-Noguer. D-NeRF: Neural Radiance Fields for Dynamic Scenes. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2021

    work page 2021

  26. [26]

    Learning transferable visual models from natural language supervision

    Alec Radford, Jong Wook Kim, Chris Hallacy, Aditya Ramesh, Gabriel Goh, Sandhini Agarwal, Girish Sastry, Amanda Askell, Pamela Mishkin, Jack Clark, et al. Learning transferable visual models from natural language supervision. InICML, 2021

    work page 2021

  27. [27]

    View-invariant representation and recognition of actions.International journal of computer vision, 50(2):203–226, 2002

    Cen Rao, Alper Yilmaz, and Mubarak Shah. View-invariant representation and recognition of actions.International journal of computer vision, 50(2):203–226, 2002

    work page 2002

  28. [28]

    L4GM: Large 4d gaussian reconstruction model

    Jiawei Ren, Kevin Xie, Ashkan Mirzaei, hanxue liang, Xiaohui Zeng, Karsten Kreis, Ziwei Liu, Antonio Torralba, Sanja Fidler, Seung Wook Kim, and Huan Ling. L4GM: Large 4d gaussian reconstruction model. InThe Thirty-eighth Annual Conference on Neural Information Processing Systems, 2024

    work page 2024

  29. [29]

    K-planes: Explicit radiance fields in space, time, and appearance

    Sara Fridovich-Keil and Giacomo Meanti, Frederik Rahbæk Warburg, Benjamin Recht, and Angjoo Kanazawa. K-planes: Explicit radiance fields in space, time, and appearance. InCVPR, 2023

    work page 2023

  30. [30]

    Convolutional wasserstein distances: Efficient optimal transportation on geometric domains.ACM Transactions on Graphics (ToG), 34(4):1–11, 2015

    Justin Solomon, Fernando De Goes, Gabriel Peyré, Marco Cuturi, Adrian Butscher, Andy Nguyen, Tao Du, and Leonidas Guibas. Convolutional wasserstein distances: Efficient optimal transportation on geometric domains.ACM Transactions on Graphics (ToG), 34(4):1–11, 2015

    work page 2015

  31. [31]

    Water-gs: Toward copyright protection for 3d gaussian splatting via universal watermarking, 2024

    Yuqi Tan, Xiang Liu, Shuzhao Xie, Bin Chen, Shu-Tao Xia, and Zhi Wang. Water-gs: Toward copyright protection for 3d gaussian splatting via universal watermarking, 2024. 12

    work page 2024

  32. [32]

    FVD: A new metric for video generation, 2019

    Thomas Unterthiner, Sjoerd van Steenkiste, Karol Kurach, Raphaël Marinier, Marcin Michalski, and Sylvain Gelly. FVD: A new metric for video generation, 2019

    work page 2019

  33. [33]

    Springer, 2009

    Cédric Villani et al.Optimal transport: old and new, volume 338. Springer, 2009

    work page 2009

  34. [34]

    Mega: Hybrid mesh-gaussian head avatar for high-fidelity rendering and head editing

    Cong Wang, Di Kang, Heyi Sun, Shenhan Qian, Zixuan Wang, Linchao Bao, and Song-Hai Zhang. Mega: Hybrid mesh-gaussian head avatar for high-fidelity rendering and head editing. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 26274–26284, June 2025

    work page 2025

  35. [35]

    Bovik, Hamid R

    Zhou Wang, Alan C. Bovik, Hamid R. Sheikh, and Eero P. Simoncelli. Image quality assessment: From error visibility to structural similarity.IEEE Transactions on Image Processing, 13(4):600– 612, 2004

    work page 2004

  36. [36]

    4d gaussian splatting for real-time dynamic scene rendering

    Guanjun Wu, Taoran Yi, Jiemin Fang, Lingxi Xie, Xiaopeng Zhang, Wei Wei, Wenyu Liu, Qi Tian, and Xinggang Wang. 4d gaussian splatting for real-time dynamic scene rendering. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2024

    work page 2024

  37. [37]

    Sc4d: Sparse- controlled video-to-4d generation and motion transfer

    Zijie Wu, Chaohui Yu, Yanqin Jiang, Chenjie Cao, Fan Wang, and Xiang Bai. Sc4d: Sparse- controlled video-to-4d generation and motion transfer. InEuropean Conference on Computer Vision, 2024

    work page 2024

  38. [38]

    Street gaussians: Modeling dynamic urban scenes with gaussian splatting

    Yunzhi Yan, Haotong Lin, Chenxu Zhou, Weijie Wang, Haiyang Sun, Kun Zhan, Xianpeng Lang, Xiaowei Zhou, and Sida Peng. Street gaussians: Modeling dynamic urban scenes with gaussian splatting. InECCV, 2024

    work page 2024

  39. [39]

    Real-time photorealistic dynamic scene representation and rendering with 4d gaussian splatting

    Zeyu Yang, Hongye Yang, Zijie Pan, and Li Zhang. Real-time photorealistic dynamic scene representation and rendering with 4d gaussian splatting. InInternational Conference on Learning Representations (ICLR), 2024

    work page 2024

  40. [40]

    Deformable 3d gaussians for high-fidelity monocular dynamic scene reconstruction.CVPR, 2023

    Ziyi Yang, Xinyu Gao, Wen Zhou, Shaohui Jiao, Yuqing Zhang, and Xiaogang Jin. Deformable 3d gaussians for high-fidelity monocular dynamic scene reconstruction.CVPR, 2023

    work page 2023

  41. [41]

    4dgen: Grounded 4d content generation with spatial-temporal consistency

    Yuyang Yin, Dejia Xu, Zhangyang Wang, Yao Zhao, and Yunchao Wei. 4dgen: Grounded 4d content generation with spatial-temporal consistency. 2023

    work page 2023

  42. [42]

    Deep 3d-to-2d watermarking: Embedding messages in 3d meshes and extracting them from 2d renderings

    Innfarn Yoo, Huiwen Chang, Xiyang Luo, Ondrej Stava, Ce Liu, Peyman Milanfar, and Feng Yang. Deep 3d-to-2d watermarking: Embedding messages in 3d meshes and extracting them from 2d renderings. In2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2022

    work page 2022

  43. [43]

    SQS: Enhancing sparse perception models via query-based splatting in autonomous driving

    Haiming Zhang, Yiyao Zhu, Wending Zhou, Xu Yan, Yingjie CAI, Bingbing Liu, Shuguang Cui, and Zhen Li. SQS: Enhancing sparse perception models via query-based splatting in autonomous driving. InThe Thirty-ninth Annual Conference on Neural Information Processing Systems, 2025

    work page 2025

  44. [44]

    4diffusion: Multi-view video diffusion model for 4d generation

    Haiyu Zhang, Xinyuan Chen, Yaohui Wang, Xihui Liu, Yunhong Wang, and Yu Qiao. 4diffusion: Multi-view video diffusion model for 4d generation. InThe Thirty-eighth Annual Conference on Neural Information Processing Systems, 2024

    work page 2024

  45. [45]

    The unreason- able effectiveness of deep features as a perceptual metric

    Richard Zhang, Phillip Isola, Alexei A Efros, Eli Shechtman, and Oliver Wang. The unreason- able effectiveness of deep features as a perceptual metric. InCVPR, 2018

    work page 2018

  46. [46]

    Gs-hider: Hiding messages into 3d gaussian splatting

    Xuanyu Zhang, Jiarui Meng, Runyi Li, Zhipei Xu, Yongbing Zhang, and Jian Zhang. Gs-hider: Hiding messages into 3d gaussian splatting. InAdvances in Neural Information Processing Systems, 2024

    work page 2024

  47. [47]

    SecureGS: Boosting the security and fidelity of 3d gaussian splatting steganography

    Xuanyu Zhang, Jiarui Meng, Zhipei Xu, Shuzhou Yang, Yanmin Wu, Ronggang Wang, and Jian Zhang. SecureGS: Boosting the security and fidelity of 3d gaussian splatting steganography. In The Thirteenth International Conference on Learning Representations, 2025

    work page 2025

  48. [48]

    Hidden: Hiding data with deep networks

    Jiren Zhu, Russell Kaplan, Justin Johnson, and Li Fei-Fei. Hidden: Hiding data with deep networks. In Vittorio Ferrari, Martial Hebert, Cristian Sminchisescu, and Yair Weiss, editors, Computer Vision – ECCV 2018, pages 682–697, Cham, 2018. Springer International Publishing. 13 A Impact Statement The rapid advancement of 4D Gaussian Splatting (4DGS) has si...

    work page 2018