pith. machine review for the scientific record. sign in

arxiv: 2604.02915 · v1 · submitted 2026-04-03 · 💻 cs.CV

Recognition: no theorem link

GP-4DGS: Probabilistic 4D Gaussian Splatting from Monocular Video via Variational Gaussian Processes

Mijeong Kim , Jungtaek Kim , Bohyung Han
Authors on Pith no claims yet

Pith reviewed 2026-05-13 20:05 UTC · model grok-4.3

classification 💻 cs.CV
keywords 4D Gaussian SplattingGaussian Processesdynamic scene reconstructionuncertainty quantificationmonocular videovariational inferencedeformation fields
0
0 comments X

The pith

GP-4DGS integrates variational Gaussian Processes into 4D Gaussian Splatting to enable probabilistic modeling of dynamic scenes from monocular video.

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

The paper introduces GP-4DGS to move 4D Gaussian Splatting from deterministic to probabilistic reconstruction of dynamic scenes. It uses Gaussian Processes to quantify uncertainty in motion predictions, estimate motion where observations are missing or sparse, and extrapolate motion beyond the training frames. To make this feasible with thousands of Gaussian primitives, the method designs spatio-temporal kernels to capture deformation correlations and applies variational inference with inducing points. Experiments indicate that this yields higher-quality reconstructions while producing uncertainty maps that highlight regions of motion ambiguity. The work connects probabilistic modeling directly to neural rendering pipelines for dynamic content.

Core claim

By embedding variational Gaussian Processes into 4D Gaussian Splatting, the approach models the deformation fields of a large set of Gaussian primitives via custom spatio-temporal kernels, supports tractable inference through inducing points, and delivers three capabilities: uncertainty quantification for predicted motions, motion estimation in unobserved regions, and temporal extrapolation past observed frames, while improving overall reconstruction fidelity.

What carries the argument

Spatio-temporal kernels inside variational Gaussian Processes that represent correlations in the deformation fields of Gaussian primitives within 4DGS.

If this is right

  • Reconstruction quality for dynamic scenes increases when probabilistic modeling replaces deterministic assumptions.
  • Uncertainty estimates reliably mark areas of high motion ambiguity in the output.
  • Motion can be estimated in regions lacking direct observations or with sparse sampling.
  • Motion predictions extend temporally beyond the input training frames.

Where Pith is reading between the lines

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

  • Robotics or augmented reality systems could use the uncertainty output to avoid acting on unreliable motion estimates.
  • The kernel-based formulation may transfer to other point-based deformation tasks beyond Gaussian Splatting.
  • Further gains in speed could come from adaptive selection of inducing points for larger scenes.

Load-bearing premise

The chosen spatio-temporal kernels sufficiently capture the correlation patterns in deformation fields across thousands of Gaussian primitives so that variational inference stays accurate.

What would settle it

Experiments where the reported uncertainty values show no correlation with actual motion prediction errors, or where reconstruction metrics fail to improve over deterministic 4DGS baselines.

Figures

Figures reproduced from arXiv: 2604.02915 by Bohyung Han, Jungtaek Kim, Mijeong Kim.

Figure 1
Figure 1. Figure 1: We propose GP-4DGS, a novel integration of Gaussian Processes (GPs) [ [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Uncertainty quantification. GP-4DGS provides princi [PITH_FULL_IMAGE:figures/full_fig_p006_2.png] view at source ↗
Figure 5
Figure 5. Figure 5: Motion extrapolation results from GP-4DGS. Our GP [PITH_FULL_IMAGE:figures/full_fig_p007_5.png] view at source ↗
Figure 4
Figure 4. Figure 4: Qualitative comparison on the DAVIS dataset under ex [PITH_FULL_IMAGE:figures/full_fig_p007_4.png] view at source ↗
Figure 8
Figure 8. Figure 8: Visualization of inducing points (paper-windmill). In￾ducing points (red) are well-distributed across the canonical space and temporal axis to ensure comprehensive coverage of the scene [PITH_FULL_IMAGE:figures/full_fig_p008_8.png] view at source ↗
Figure 6
Figure 6. Figure 6: Trajectory comparison on the (left) paper-windmill and (right) block scene. GP guidance effectively regularizes motion trajectories, reducing noise and producing physically plausible motion patterns, compared to the baseline approach. 0.0 0.2 0.4 0.6 0.8 1.0 Time (0-1) 1 0 1 2 Deformation (a) Trajectory w/o GP 0.0 0.2 0.4 0.6 0.8 1.0 Time (0-1) 1.5 1.0 0.5 0.0 0.5 1.0 1.5 Deformation (b) Trajectory from GP… view at source ↗
Figure 7
Figure 7. Figure 7: Trajectory comparison on the spin scene between the initial GS reconstruction and the GP guidance in the GP-GS op￾timization. These graphs correspond to the first dimension in 6D rotation. The GP provides accurate and stable motion priors. ability, we adopt the Area Under the Sparsification Error (AUSE), which measures the alignment between estimated uncertainty and actual reconstruction error. As shown in… view at source ↗
read the original abstract

We present GP-4DGS, a novel framework that integrates Gaussian Processes (GPs) into 4D Gaussian Splatting (4DGS) for principled probabilistic modeling of dynamic scenes. While existing 4DGS methods focus on deterministic reconstruction, they are inherently limited in capturing motion ambiguity and lack mechanisms to assess prediction reliability. By leveraging the kernel-based probabilistic nature of GPs, our approach introduces three key capabilities: (i) uncertainty quantification for motion predictions, (ii) motion estimation for unobserved or sparsely sampled regions, and (iii) temporal extrapolation beyond observed training frames. To scale GPs to the large number of Gaussian primitives in 4DGS, we design spatio-temporal kernels that capture the correlation structure of deformation fields and adopt variational Gaussian Processes with inducing points for tractable inference. Our experiments show that GP-4DGS enhances reconstruction quality while providing reliable uncertainty estimates that effectively identify regions of high motion ambiguity. By addressing these challenges, our work takes a meaningful step toward bridging probabilistic modeling and neural graphics.

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 paper proposes GP-4DGS, a framework that augments 4D Gaussian Splatting with variational Gaussian Processes using custom spatio-temporal kernels and inducing-point approximations. It claims this enables three new capabilities for dynamic scene reconstruction from monocular video: calibrated uncertainty quantification over motion, inpainting of deformation fields in unobserved or sparsely sampled regions, and temporal extrapolation beyond the training frames, while maintaining or improving reconstruction fidelity.

Significance. If the central claims are substantiated, the work would meaningfully extend deterministic 4DGS methods by introducing principled probabilistic modeling. The ability to report motion uncertainty and perform extrapolation could benefit downstream tasks such as video prediction and robotics. The scalability approach via inducing points addresses a practical barrier in applying GPs to the high-dimensional per-primitive deformation fields typical of 4DGS.

major comments (2)
  1. [Abstract] Abstract: The abstract asserts that experiments demonstrate enhanced reconstruction quality and reliable uncertainty estimates that identify regions of high motion ambiguity, yet no quantitative metrics, baselines, error tables, or held-out frame evaluations are referenced. Without these, the three advertised capabilities cannot be assessed.
  2. [Method/Experiments] Method and Experiments: The central claim hinges on the spatio-temporal kernels (product or sum of spatial RBF/SE and temporal kernels) plus variational inference with inducing points preserving deformation fidelity for O(10^4–10^5) Gaussians. The manuscript must show that the optimized ELBO produces posterior variances that correlate with actual reconstruction error on held-out frames and that reducing the number of inducing points does not materially degrade uncertainty calibration or extrapolation performance; absent such verification, the capabilities reduce to the quality of the mean function alone.
minor comments (2)
  1. [Method] Clarify the exact form of the spatio-temporal kernel (additive vs. multiplicative) and list all free hyperparameters and inducing-point placement strategy in a dedicated subsection.
  2. [Experiments] Add a table or figure comparing reconstruction metrics (PSNR, SSIM, LPIPS) against standard 4DGS baselines on the same datasets to quantify the claimed enhancement.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the positive evaluation of our work's significance and for the constructive comments. We have carefully addressed each major comment and revised the manuscript accordingly.

read point-by-point responses

  1. Referee: [Abstract] Abstract: The abstract asserts that experiments demonstrate enhanced reconstruction quality and reliable uncertainty estimates that identify regions of high motion ambiguity, yet no quantitative metrics, baselines, error tables, or held-out frame evaluations are referenced. Without these, the three advertised capabilities cannot be assessed.

    Authors: We agree that the abstract would benefit from referencing key quantitative aspects to better support the claims. In the revised manuscript, we have updated the abstract to briefly mention the use of metrics such as PSNR and SSIM for reconstruction quality, as well as calibration measures for uncertainty on held-out frames. The detailed tables, baselines, and evaluations are already present in the Experiments section (Section 4), including comparisons with deterministic 4DGS methods. revision: yes

  2. Referee: [Method/Experiments] Method and Experiments: The central claim hinges on the spatio-temporal kernels (product or sum of spatial RBF/SE and temporal kernels) plus variational inference with inducing points preserving deformation fidelity for O(10^4–10^5) Gaussians. The manuscript must show that the optimized ELBO produces posterior variances that correlate with actual reconstruction error on held-out frames and that reducing the number of inducing points does not materially degrade uncertainty calibration or extrapolation performance; absent such verification, the capabilities reduce to the quality of the mean function alone.

    Authors: This is a valid point regarding the need for explicit verification of the uncertainty estimates. The original manuscript includes qualitative evidence of uncertainty correlating with motion ambiguity and quantitative reconstruction results. To strengthen this, we have added in the revision: correlation plots between posterior variance and held-out reconstruction error, and an ablation study on the number of inducing points showing minimal degradation in calibration (via expected calibration error) and extrapolation performance for the tested range. These confirm the variational GP approximation maintains the probabilistic modeling benefits. revision: yes

Circularity Check

0 steps flagged

No circularity: derivation relies on standard variational GP machinery applied to 4DGS deformations

full rationale

The paper applies established variational Gaussian process inference (inducing-point ELBO) together with product/sum spatio-temporal kernels to per-primitive deformation fields. The three advertised capabilities—uncertainty, inpainting of unobserved regions, and extrapolation—follow directly from the GP posterior once the kernel and variational approximation are chosen; no equation reduces a claimed prediction to a fitted parameter by construction, no uniqueness theorem is imported from the authors' prior work, and no ansatz is smuggled via self-citation. The central modeling choice (kernel design + inducing points) is justified by tractability arguments that remain independent of the final reconstruction metrics.

Axiom & Free-Parameter Ledger

2 free parameters · 2 axioms · 0 invented entities

Central claim rests on standard assumptions from Gaussian Process literature and 4DGS deformation modeling; no invented entities are introduced. Free parameters include kernel hyperparameters and variational parameters for inducing points, which are fitted during training.

free parameters (2)
  • spatio-temporal kernel hyperparameters
    Parameters of the kernels that capture deformation correlations are chosen or optimized to fit the data.
  • number and placement of inducing points
    Inducing points for variational approximation are selected to scale inference to many Gaussians.
axioms (2)
  • domain assumption Gaussian Processes with spatio-temporal kernels can model the deformation fields of 4D Gaussian primitives
    Invoked when stating that GPs capture motion correlation structure for uncertainty quantification.
  • standard math Variational inference with inducing points yields tractable and accurate approximations for large-scale GP regression
    Standard assumption from variational GP literature used to justify scalability.

pith-pipeline@v0.9.0 · 5488 in / 1359 out tokens · 34865 ms · 2026-05-13T20:05:48.493671+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

55 extracted references · 55 canonical work pages · 1 internal anchor

  1. [1]

    Mad- dix, Michael W

    Abdul Fatir Ansari, Lorenzo Stella, Caner Turkmen, Xiyuan Zhang, Pedro Mercado, Huibin Shen, Oleksandr Shchur, Syama Syndar Rangapuram, Sebastian Pineda Arango, Shubham Kapoor, Jasper Zschiegner, Danielle C. Mad- dix, Michael W. Mahoney, Kari Torkkola, Andrew Gor- don Wilson, Michael Bohlke-Schneider, and Yuyang Wang. Chronos: Learning the language of tim...

    work page 2024

  2. [2]

    DynaSurfGS: Dynamic surface reconstruction with planar- based Gaussian splatting.arXiv preprint arXiv:2408.13972,

    Weiwei Cai, Weicai Ye, Peng Ye, Tong He, and Tao Chen. DynaSurfGS: Dynamic surface reconstruction with planar- based Gaussian splatting.arXiv preprint arXiv:2408.13972,

    work page arXiv

  3. [3]

    TensoRF: Tensorial radiance fields

    Anpei Chen, Zexiang Xu, Andreas Geiger, Jingyi Yu, and Hao Su. TensoRF: Tensorial radiance fields. InECCV, 2022. 1

    work page 2022

  4. [4]

    Neural radiance flow for 4D view synthesis and video processing

    Yilun Du, Yinan Zhang, Hong-Xing Yu, Joshua B Tenen- baum, and Jiajun Wu. Neural radiance flow for 4D view synthesis and video processing. InICCV, 2021. 2

    work page 2021

  5. [5]

    4d-rotor gaussian splatting: towards efficient novel view synthesis for dynamic scenes

    Yuanxing Duan, Fangyin Wei, Qiyu Dai, Yuhang He, Wen- zheng Chen, and Baoquan Chen. 4d-rotor gaussian splatting: towards efficient novel view synthesis for dynamic scenes. In SIGGRAPH, 2024. 2

    work page 2024

  6. [6]

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

    Sara Fridovich-Keil, Giacomo Meanti, Frederik Rahbæk Warburg, Benjamin Recht, and Angjoo Kanazawa. K-planes: Explicit radiance fields in space, time, and appearance. In CVPR, 2023. 2

    work page 2023

  7. [7]

    Dynamic view synthesis from dynamic monocular video

    Chen Gao, Ayush Saraf, Johannes Kopf, and Jia-Bin Huang. Dynamic view synthesis from dynamic monocular video. In ICCV, 2021. 2

    work page 2021

  8. [8]

    Monocular dynamic view synthesis: A reality check

    Hang Gao, Ruilong Li, Shubham Tulsiani, Bryan Russell, and Angjoo Kanazawa. Monocular dynamic view synthesis: A reality check. InNeurIPS, 2022. 6

    work page 2022

  9. [9]

    FastNeRF: High-fidelity neural rendering at 200FPS

    Stephan J Garbin, Marek Kowalski, Matthew Johnson, Jamie Shotton, and Julien Valentin. FastNeRF: High-fidelity neural rendering at 200FPS. InICCV, 2021. 1

    work page 2021

  10. [10]

    GPyTorch: Blackbox matrix- matrix Gaussian process inference with GPU acceleration

    Jacob Gardner, Geoff Pleiss, Kilian Q Weinberger, David Bindel, and Andrew G Wilson. GPyTorch: Blackbox matrix- matrix Gaussian process inference with GPU acceleration. In NIPS, 2018. 6

    work page 2018

  11. [11]

    Motion-aware 3d gaussian splatting for effi- cient dynamic scene reconstruction

    Zhiyang Guo, Wengang Zhou, Li Li, Min Wang, and Houqiang Li. Motion-aware 3d gaussian splatting for effi- cient dynamic scene reconstruction. InTCSVT, 2024. 2

    work page 2024

  12. [12]

    GP-GS: Gaussian processes for enhanced Gaussian splatting.arXiv preprint arXiv:2502.02283, 2025

    Zhihao Guo, Jingxuan Su, Shenglin Wang, Jinlong Fan, Jing Zhang, Liangxiu Han, and Peng Wang. GP-GS: Gaussian processes for enhanced Gaussian splatting.arXiv preprint arXiv:2502.02283, 2025. 2, 3

    work page arXiv 2025

  13. [13]

    LRM: Large reconstruction model for single image to 3D

    Yicong Hong, Kai Zhang, Jiuxiang Gu, Sai Bi, Yang Zhou, Difan Liu, Feng Liu, Kalyan Sunkavalli, Trung Bui, and Hao Tan. LRM: Large reconstruction model for single image to 3D. InICLR, 2024. 1

    work page 2024

  14. [14]

    Vivid4d: Improving 4d reconstruction from monocular video by video inpainting

    Jiaxin Huang, Sheng Miao, Bangbang Yang, Yuewen Ma, and Yiyi Liao. Vivid4d: Improving 4d reconstruction from monocular video by video inpainting. InICCV, 2025. 6, 7, 12

    work page 2025

  15. [15]

    SC-GS: Sparse-controlled Gaussian splatting for editable dynamic scenes

    Yi-Hua Huang, Yang-Tian Sun, Ziyi Yang, Xiaoyang Lyu, Yan-Pei Cao, and Xiaojuan Qi. SC-GS: Sparse-controlled Gaussian splatting for editable dynamic scenes. InCVPR,

    work page

  16. [16]

    3D Gaussian splatting for real-time radiance field rendering

    Bernhard Kerbl, Georgios Kopanas, Thomas Leimk ¨uhler, and George Drettakis. 3D Gaussian splatting for real-time radiance field rendering. InACM TOG, 2023. 1, 2

    work page 2023

  17. [17]

    3D Gaussian splatting as Markov Chain Monte Carlo

    Shakiba Kheradmand, Daniel Rebain, Gopal Sharma, Wei- wei Sun, Yang-Che Tseng, Hossam Isack, Abhishek Kar, Andrea Tagliasacchi, and Kwang Moo Yi. 3D Gaussian splatting as Markov Chain Monte Carlo. InNeurIPS, 2024. 2, 3

    work page 2024

  18. [18]

    InfoNeRF: Ray entropy minimization for few-shot neural volume ren- dering

    Mijeong Kim, Seonguk Seo, and Bohyung Han. InfoNeRF: Ray entropy minimization for few-shot neural volume ren- dering. InCVPR, 2022. 1

    work page 2022

  19. [19]

    4D Gaus- sian splatting in the wild with uncertainty-aware regulariza- tion

    Mijeong Kim, Jongwoo Lim, and Bohyung Han. 4D Gaus- sian splatting in the wild with uncertainty-aware regulariza- tion. InNeurIPS, 2024. 2, 8

    work page 2024

  20. [20]

    DynMF: Neural motion factorization for real-time dynamic view synthesis with 3D Gaussian splatting

    Agelos Kratimenos, Jiahui Lei, and Kostas Daniilidis. DynMF: Neural motion factorization for real-time dynamic view synthesis with 3D Gaussian splatting. InECCV, 2024. 2

    work page 2024

  21. [21]

    Mosca: Dynamic gaussian fusion from casual videos via 4d motion scaffolds.arXiv preprint arXiv:2405.17421, 2024

    Jiahui Lei, Yijia Weng, Adam Harley, Leonidas Guibas, and Kostas Daniilidis. Mosca: Dynamic gaussian fusion from casual videos via 4d motion scaffolds, 2024. arXiv preprint arXiv:2405.17421. 2

    work page arXiv 2024

  22. [22]

    DynIBaR: Neural dynamic image-based rendering

    Zhengqi Li, Qianqian Wang, Forrester Cole, Richard Tucker, and Noah Snavely. DynIBaR: Neural dynamic image-based rendering. InCVPR, 2023. 6

    work page 2023

  23. [23]

    Spacetime Gaussian feature splatting for real-time dynamic view syn- thesis

    Zhan Li, Zhang Chen, Zhong Li, and Yi Xu. Spacetime Gaussian feature splatting for real-time dynamic view syn- thesis. InCVPR, 2024. 2, 3, 6

    work page 2024

  24. [24]

    MegaSaM: Accurate, fast and robust structure and motion from casual dynamic videos

    Zhengqi Li, Richard Tucker, Forrester Cole, Qianqian Wang, Linyi Jin, Vickie Ye, Angjoo Kanazawa, Aleksander Holyn- ski, and Noah Snavely. MegaSaM: Accurate, fast and robust structure and motion from casual dynamic videos. InCVPR,

    work page

  25. [25]

    Gaufre: Gaussian deformation fields for real-time dynamic novel view synthesis

    Yiqing Liang, Numair Khan, Zhengqin Li, Thu Nguyen- Phuoc, Douglas Lanman, James Tompkin, and Lei Xiao. Gaufre: Gaussian deformation fields for real-time dynamic novel view synthesis. InWACV, 2025. 2

    work page 2025

  26. [26]

    Gaussian-flow: 4d reconstruction with dynamic 3d gaussian particle

    Youtian Lin, Zuozhuo Dai, Siyu Zhu, and Yao Yao. Gaussian-flow: 4d reconstruction with dynamic 3d gaussian particle. InCVPR, 2024

    work page 2024

  27. [27]

    Modgs: Dy- namic gaussian splatting from casually-captured monocular videos

    Qingming Liu, Yuan Liu, Jiepeng Wang, Xianqiang Lyv, Peng Wang, Wenping Wang, and Junhui Hou. Modgs: Dy- namic gaussian splatting from casually-captured monocular videos. InICLR, 2025. 2

    work page 2025

  28. [28]

    VL-Grasp: a 6-Dof interactive grasp policy for language-oriented objects in cluttered indoor scenes

    Yuhao Lu, Yixuan Fan, Beixing Deng, Fangfu Liu, Yali Li, and Shengjin Wang. VL-Grasp: a 6-Dof interactive grasp policy for language-oriented objects in cluttered indoor scenes. InIROS, 2023. 1

    work page 2023

  29. [29]

    3d geometry-aware deformable gaussian splatting for dynamic view synthesis

    Zhicheng Lu, Xiang Guo, Le Hui, Tianrui Chen, Min Yang, Xiao Tang, Feng Zhu, and Yuchao Dai. 3d geometry-aware deformable gaussian splatting for dynamic view synthesis. InCVPR, 2024. 2

    work page 2024

  30. [30]

    Srinivasan, Matthew Tancik, Jonathan T

    Ben Mildenhall, Pratul P. Srinivasan, Matthew Tancik, Jonathan T. Barron, Ravi Ramamoorthi, and Ren Ng. NeRF: Representing scenes as neural radiance fields for view syn- thesis. InECCV, 2020. 1, 2

    work page 2020

  31. [31]

    DriveWorld: 4D pre-trained scene understanding via world models for au- tonomous driving

    Chen Min, Dawei Zhao, Liang Xiao, Jian Zhao, Xinli Xu, Zheng Zhu, Lei Jin, Jianshu Li, Yulan Guo, Junliang Xing, Liping Jing, Yiming Nie, and Bin Dai. DriveWorld: 4D pre-trained scene understanding via world models for au- tonomous driving. InCVPR, 2024. 1

    work page 2024

  32. [32]

    Instant neural graphics primitives with a multires- olution hash encoding.ACM TOG, 2022

    Thomas M ¨uller, Alex Evans, Christoph Schied, and Alexan- der Keller. Instant neural graphics primitives with a multires- olution hash encoding.ACM TOG, 2022. 1

    work page 2022

  33. [33]

    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. 2

    work page 2021

  34. [34]

    Barron, Sofien Bouaziz, Dan B Goldman, Ricardo Martin- Brualla, and Steven M

    Keunhong Park, Utkarsh Sinha, Peter Hedman, Jonathan T. Barron, Sofien Bouaziz, Dan B Goldman, Ricardo Martin- Brualla, and Steven M. Seitz. HyperNeRF: A higher- dimensional representation for topologically varying neural radiance fields.ACM TOG, 40(6):1–12, 2021. 6

    work page 2021

  35. [35]

    DeSiRe-GS: 4D street gaussians for static- dynamic decomposition and surface reconstruction for urban driving scenes

    Yuxin Peng et al. DeSiRe-GS: 4D street gaussians for static- dynamic decomposition and surface reconstruction for urban driving scenes. InCVPR, 2025. 1

    work page 2025

  36. [36]

    The 2017 DAVIS Challenge on Video Object Segmentation

    Jordi Pont-Tuset, Federico Perazzi, Sergi Caelles, Pablo Ar- bel´aez, Alex Sorkine-Hornung, and Luc Van Gool. The 2017 DA VIS challenge on video object segmentation.arXiv preprint arXiv:1704.00675, 2017. 6

    work page internal anchor Pith review arXiv 2017

  37. [37]

    Gaussian Processes for Machine Learning

    Carl Edward Rasmussen and Christopher KI Williams. Gaussian Processes for Machine Learning. MIT Press, 2006. 1, 2

    work page 2006

  38. [38]

    Rig3DGS: Creating controllable portraits from ca- sual monocular videos.arXiv:2402.03723, 2024

    Alfredo Rivero, ShahRukh Athar, Zhixin Shu, and Dimitris Samaras. Rig3DGS: Creating controllable portraits from ca- sual monocular videos.arXiv:2402.03723, 2024. 2

    work page arXiv 2024

  39. [39]

    NeRF- SLAM: Real-time dense monocular SLAM with neural radi- ance fields

    Antoni Rosinol, John J Leonard, and Luca Carlone. NeRF- SLAM: Real-time dense monocular SLAM with neural radi- ance fields. InIROS, 2023. 1

    work page 2023

  40. [40]

    Mod- eling uncertainty for Gaussian splatting.arXiv preprint arXiv:2403.18476, 2024

    Luca Savant, Diego Valsesia, and Enrico Magli. Mod- eling uncertainty for Gaussian splatting.arXiv preprint arXiv:2403.18476, 2024. 2, 8

    work page arXiv 2024

  41. [41]

    Tensor4D: Efficient neural 4D decomposition for high-fidelity dynamic reconstruction and rendering

    Ruizhi Shao, Zerong Zheng, Hanzhang Tu, Boning Liu, Hongwen Zhang, and Yebin Liu. Tensor4D: Efficient neural 4D decomposition for high-fidelity dynamic reconstruction and rendering. InCVPR, 2023. 2

    work page 2023

  42. [42]

    Dynamic gaussian marbles for novel view synthesis of casual monocular videos

    Colton Stearns, Adam Harley, Mikaela Uy, Florian Dubost, Federico Tombari, Gordon Wetzstein, and Leonidas Guibas. Dynamic gaussian marbles for novel view synthesis of casual monocular videos. InSIGGRAPH, 2024. 2, 6, 7, 13

    work page 2024

  43. [43]

    Variational learning of inducing variables in sparse Gaussian processes

    Michalis K Titsias. Variational learning of inducing variables in sparse Gaussian processes. InAISTATS, 2009. 4

    work page 2009

  44. [44]

    Variational bayes gaussian splatting

    Toon Van de Maele, Ozan Catal, Alexander Tschantz, Christopher L Buckley, and Tim Verbelen. Variational bayes gaussian splatting. InNeurIPS, 2025. 2, 3

    work page 2025

  45. [45]

    D-miso: Editing dynamic 3d scenes using multi-gaussians soup

    Joanna Waczynska, Piotr Borycki, Joanna Kaleta, Slawomir Tadeja, and Przemysław Spurek. D-miso: Editing dynamic 3d scenes using multi-gaussians soup. InNeurIPS, 2024. 2

    work page 2024

  46. [46]

    Shape of Mo- tion: 4D reconstruction from a single video

    Qianqian Wang, Vickie Ye, Hang Gao, Weijia Zeng, Jake Austin, Zhengqi Li, and Angjoo Kanazawa. Shape of Mo- tion: 4D reconstruction from a single video. InICCV, 2025. 2, 3, 6, 7, 13

    work page 2025

  47. [47]

    Gflow: Recovering 4D world from monocular video

    Shizun Wang, Xingyi Yang, Qiuhong Shen, Zhenxiang Jiang, and Xinchao Wang. Gflow: Recovering 4D world from monocular video. InAAAI, 2025

    work page 2025

  48. [48]

    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. InCVPR, 2024. 2, 3, 6

    work page 2024

  49. [49]

    Deblur4DGS: 4D Gaussian splatting from blurry monocular video.arXiv preprint arXiv:2412.06424,

    Renlong Wu, Zhilu Zhang, Mingyang Chen, Zifei Yan, and Wangmeng Zuo. Deblur4DGS: 4D Gaussian splatting from blurry monocular video.arXiv preprint arXiv:2412.06424,

    work page arXiv

  50. [50]

    DrivingSphere: Building a high-fidelity 4D world for closed- loop simulation

    Tianyi Yan, Dongming Wu, Wencheng Han, Junpeng Jiang, Xia Zhou, Kun Zhan, Cheng-zhong Xu, and Jianbing Shen. DrivingSphere: Building a high-fidelity 4D world for closed- loop simulation. InCVPR, 2025. 1

    work page 2025

  51. [51]

    Deformable 3D Gaussians for high-fidelity monocular dynamic scene reconstruction

    Ziyi Yang, Xinyu Gao, Wen Zhou, Shaohui Jiao, Yuqing Zhang, and Xiaogang Jin. Deformable 3D Gaussians for high-fidelity monocular dynamic scene reconstruction. In CVPR, 2024. 2, 3, 6

    work page 2024

  52. [52]

    Online learning of neural surface light fields alongside real-time incremental 3D re- construction.IEEE Robotics and Automation Letters, 8(6): 3844–3851, 2023

    Yijun Yuan and Andreas N ¨uchter. Online learning of neural surface light fields alongside real-time incremental 3D re- construction.IEEE Robotics and Automation Letters, 8(6): 3844–3851, 2023. 1

    work page 2023

  53. [53]

    GeoLRM: Geometry-aware large recon- struction model for high-quality 3D Gaussian generation

    Chubin Zhang, Hongliang Song, Yi Wei, Chen Yu, Jiwen Lu, and Yansong Tang. GeoLRM: Geometry-aware large recon- struction model for high-quality 3D Gaussian generation. In NeuIPS, 2024. 1

    work page 2024

  54. [54]

    DrivingGaussian: Composite gaussian splatting for surrounding dynamic au- tonomous driving scenes

    Xiaoyu Zhou, Zhiwei Lin, Xiaojun Shan, Yongtao Wang, Deqing Sun, and Ming-Hsuan Yang. DrivingGaussian: Composite gaussian splatting for surrounding dynamic au- tonomous driving scenes. InCVPR, 2024. 1

    work page 2024

  55. [55]

    On the continuity of rotation representations in neural networks

    Yi Zhou, Connelly Barnes, Jingwan Lu, Jimei Yang, and Hao Li. On the continuity of rotation representations in neural networks. InCVPR, 2019. 4 A. Effect of Time-series Feature Extractor MotivationIn Variational Gaussian Processes (VGPs), the initialization of inducing pointsZis a key factor in summarizing the 4D deformation field. To represent com- plex ...

    work page 2019