IGFuse: Interactive 3D Gaussian Scene Reconstruction via Multi-Scans Fusion

Gaoang Wang; Haonan Zhou; Liu Liu; Wenhao Hu; Xi Li; Xuexiang Wen; Zesheng Li; Zhizhong Su

REVIEW 4 major objections 4 minor 41 references

IGFuse: Interactive 3D Gaussian Scene Reconstruction via Multi-Scans Fusion

T0 review · 4 major / 4 minor · reviewed 2026-08-15 · deepseek-v4-flash

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

Pith's one-line read IGFuse claims that fusing a few scans of a rearranged scene, through bi-directional photometric and semantic consistency with pseudo-intermediate alignment, reconstructs complete manipulable Gaussian scenes without inpainting or dense…

desk verdict IGFuse is a genuinely useful fusion idea under known object rearrangements, but the end-to-end claim outruns an input assumption that carries the method. read the letter →

arxiv 2508.13153 v1 pith:IPKZKD2L submitted 2025-08-18 cs.CV

Wenhao Hu , Zesheng Li , Haonan Zhou , Liu Liu , Xuexiang Wen , Zhizhong Su , Xi Li , Gaoang Wang This is my paper · ORCID

classification cs.CV

keywords 3DGaussiansplattingmulti-scanfusioninteractivescenereconstructionsegmentationnovelstatesynthesisobject-levelmanipulationreal-to-simtransferview

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

This paper tries to establish that multi-scan fusion, rather than inpainting or per-object dense scanning, is the right route to complete interactive 3D scene reconstruction. It argues that when objects move between captures, the rearrangement reveals previously occluded surfaces, so each scan can supervise and complete the others. The method builds a segmentation-aware Gaussian field for each scan, aligns scan pairs through a pseudo-intermediate scene state, and enforces reciprocal photometric and semantic consistency while co-pruning inconsistent Gaussians. If the reported results hold, a few casually captured scans would suffice for rendering and object-level manipulation, which matters for robotics simulators and real-to-simulation transfer.

What carries the argument

The load-bearing mechanism is pair-wise joint optimization of two segmentation-aware 3D Gaussian fields with reciprocal supervision: after alignment by known rigid object transformations, each scan's photometric and semantic renderings supervise the other, so surfaces occluded in one scan are completed by the other. A pseudo-intermediate scene state adds interpolation supervision that stabilizes alignment across configurations, and collaborative co-pruning removes Gaussians that remain inconsistent after alignment. Object-level manipulation is then produced by applying the known per-object rigid transformation to foreground Gaussians while leaving the background fixed.

What would settle it

Withhold the provided object poses and masks on the real-world split, estimate them by standard segmentation and registration, and measure novel-state PSNR; if it falls to the 21-22 dB single-scan baseline level, the claimed fusion depends on the privileged inputs rather than the method alone.

Watch

Extended reading notes

Core claim

The authors claim that an interactive 3D Gaussian scene can be reconstructed end-to-end by jointly optimizing one segmentation-aware Gaussian field per input scan, under bi-directional photometric and semantic consistency, with a pseudo-intermediate scene state providing unified alignment and collaborative co-pruning refining geometry. On this basis, moving objects to novel configurations yields plausible consistent renderings without dense observations or multi-stage inpainting pipelines. The paper reports novel-state synthesis quality of 36.93 PSNR on synthetic data and 27.18 PSNR on real data, against 28-30 and 21-22 PSNR for single-scan baselines, and a 4-point mIoU gain in segmentation.

Load-bearing premise

The whole reconstruction assumes every scan already comes with correct per-object rigid transformations and segmentation masks, and no image-based estimator for these inputs is supplied in the paper.

Editorial extensions

If this is right

Interaction-driven multi-scan capture can replace inpainting for completing occluded backgrounds and objects in Gaussian scenes.
Moving objects to arbitrary novel configurations renders plausibly and consistently, supporting object-level scene editing without dense per-object capture.
Cross-scan mutual supervision improves segmentation accuracy at object boundaries, reported as 91.0 mIoU versus 86.8 for a single-scan baseline.
The few-scan, no-inpainting setting lowers the capture cost for real-to-simulation transfer used in robotics training.
Simply adding more scan data to single-scan baselines yields little gain; the consistency and alignment mechanism, not the extra data alone, drives the improvement.

Reading between the lines

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

If the alignment assumption were relaxed by estimating object poses and masks jointly during fusion, the same reciprocal-supervision principle could extend to fully automatic capture.
The bi-directional consistency idea is portable to other scene representations, such as neural radiance fields or meshes, wherever multiple partial observations of a rearranged scene are available.
A direct extension would be measuring how novel-state PSNR degrades as the number of scans shrinks, which would reveal the minimum rearrangement needed for complete coverage.

Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Desk Editor's Note

read the letter

Short version: this is a real contribution with a load-bearing assumption the paper under-sells. What's new is treating multiple scans of the same scene under natural object rearrangements as mutually completing observations, then jointly optimizing segmentation-aware Gaussian fields with bi-directional photometric and semantic consistency. The pseudo-intermediate Gaussian state and collaborative co-pruning are plausible mechanisms, and the ablation attributes most of the gain to the pseudo-state (35.10 to 36.93 PSNR). The appendix's fair comparison, giving single-scan baselines the same multi-scan training data, is the right check, and those baselines don't improve much. So the headline numbers (36.93 synthetic, 27.18 real) aren't just from feeding in more data.

The soft spot is exactly what the stress-test says. The method assumes known rigid per-object transformations and input segmentation masks; Eq. (3)-(5) make the synthesized state explicit functions of those annotations. The paper calls itself end-to-end and says it avoids complex pipelines, but it never says where these inputs come from in real scenes, reports no misalignment or mask-error sensitivity, and provides no estimator. If the real-data results are obtained with externally tracked or manually specified transforms, the comparison is against an oracle alignment, and the claim that this works on unannotated captures is unsupported. That doesn't kill the paper: fusion under known rearrangements is still a useful capability, e.g., for robotic simulators where object poses are tracked. But the authors need to scope the claim.

Secondary issues: the pseudo-intermediate state and co-pruning are not formally specified; the ablation is only three rows; no error bars or repeated-run statistics; no code or data released. None of these undermines the central idea, but they make the large reported gains uncheckable from the text. The citation pattern is fine; the related work is relevant and not padded.

Bottom line: this deserves a serious referee. The novelty is real, and the core mechanism is coherent. A referee should push for code and data, a formal description of the pseudo-state, and sensitivity analysis on transformations and mask noise. If those come through, this could be a solid contribution; right now it is conditional, not broken.

Referee Report

4 major / 4 minor

Summary. The manuscript proposes IGFuse, a framework that reconstructs a complete and interactive 3D Gaussian scene by fusing several scans of the same scene captured under different object arrangements. Each scan is represented by a segmentation-aware Gaussian field, and scan pairs are aligned under assumed known rigid object transformations while enforcing bidirectional photometric and semantic consistency. A pseudo-intermediate Gaussian state and a collaborative co-pruning procedure are introduced to handle misalignment and remove inconsistent Gaussians. The authors evaluate novel-state synthesis on synthetic and real datasets, reporting PSNR 36.93 and 27.18 with large margins over single-scan baselines, and they provide additional segmentation-quality and object-completeness analyses.

Significance. If the claims hold under realistic input conditions, the paper would offer a useful alternative to inpainting-based or dense-capture pipelines for building manipulable scene representations: natural rearrangements between captures provide occlusion supervision, and the fused representation supports direct object-level transformation. Strengths include the clear multi-scan idea, bi-directional consistency losses, ablations isolating the pseudo-state contribution, and the fair-comparison experiments in Appendix H showing that merely training single-scan methods on additional scans does not close the gap. The main caveat is that the current manuscript measures performance conditional on known transforms and masks and does not specify or test the estimation of these quantities, so the headline 'end-to-end' capability is not yet established.

major comments (4)

[Section 3.2, Eqs. (2)-(5)] The entire fusion and novel-state synthesis pipeline assumes known rigid object transformations and per-scan segmentation masks as inputs: Eq. (5) makes each Gaussian's motion a function of the semantic label o_i and the transformation T^(o), and Section 3.2 states 'Using known rigid object transformations, we align the pair.' The manuscript does not describe how these quantities are obtained for real scans and provides no sensitivity analysis under misalignment or mask errors. Since the real-world PSNR 27.18 and the 'end-to-end' claim in the abstract and Figure 2 are predicated on this oracle input, the claimed capability is not demonstrated for unannotated captures. The authors should either supply a pose/mask estimation module or explicitly reframe the contribution as conditional on known annotations, and they should report robustness experiments.
[Section 3.2, pseudo-intermediate state] The pseudo-intermediate Gaussian state is introduced as a central novelty, and Table 3 attributes a large PSNR improvement to it (35.55 to 36.93), but no definition, construction rule, optimization objective, or geometric relation to the two scan states is given anywhere in the supplied text. This makes the main contribution non-reproducible. Please provide a formal specification, including how pseudo-states are initialized and optimized and how the bidirectional losses are applied to them.
[Section 3.2, Eq. (2)] The joint objective L_joint is never expanded. The 'bi-directional photometric and semantic consistency' losses and the 'collaborative co-pruning' mechanism are described only verbally, without loss terms, thresholds, schedules, or an algorithm. As a result, the method cannot be reimplemented from the manuscript and the ablation in Table 3 cannot be reproduced. Add a complete method section with explicit equations for the photometric loss, the semantic feature loss, and the co-pruning criterion.
[Tables 1-3 and 5-6] No error bars, repeated-run statistics, or significance measures are reported for any quantitative result. The claimed gains are large, but with only a small number of scenes (the dataset and scene counts are not stated in the supplied text), the stability of PSNR 36.93/27.18 across random seeds, scan-pair choices, and initialization is unknown. Report mean and standard deviation over at least three runs and state the number of scenes and views.

minor comments (4)

[Section 4.5] The section title 'Dual Guassian Convergence' contains a typo and should read 'Dual Gaussian Convergence'; the section describes discrepancies between fields but does not present quantitative convergence curves or a figure, so the reader cannot assess the claimed convergence behavior.
[Section 3.2] The terms 'pseudo scene state' and 'pseudo-intermediate Gaussian state' are used interchangeably; please choose one definition and use it consistently throughout.
[Eq. (1)] The rendered segmentation feature S is not connected to the final mask prediction or to the semantic consistency loss; please explain how S is converted to masks and how the semantic loss is computed.
[Appendix H] The text says 'incorporating multi-scans brings only slight improvements' for Table 6, but the table caption and surrounding discussion should clarify whether the baseline numbers are re-run under identical training schedules; otherwise the 'fair comparison' claim is hard to verify.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: novel-state synthesis is evaluated conditionally on known inputs (transformations and masks), and the central fusion result is benchmarked against ground-truth images and external baselines.

full rationale

The paper's central claim is empirical: fusing multi-scan observations with bi-directional photometric/semantic consistency, a pseudo-intermediate state, and co-pruning should improve rendering of novel object configurations. The evaluation compares rendered images against ground-truth target images and against external single-scan methods (GaussianEditor, Gaussian Grouping, GaussianCut, DecoupledGaussian), so the PSNR/SSIM numbers are not derived from the method's own loss by construction. The known rigid transformations T and masks M_i are inputs, not fitted parameters: Eq. (3) applies T to the optimized Gaussians to synthesize a new state, which is the intended conditional operation (analogous to providing a camera pose for novel-view synthesis), and the measured quality still depends on the learned representation. The mask-conditioned segmentation comparison in Table 5 is a reconstruction/self-consistency check rather than an independent prediction, and it is not load-bearing for the main fusion claim. The only self-citations appear in the related-work survey (Hu et al. 2025, 2024) and are not used to justify the method's correctness. No step in the derivation chain reduces to its own inputs by definition.

Assumptions & free parameters 1 free parameters · 4 assumptions · 1 invented entities

The central method rests on external per-scan supervision (masks and object transformations), a rigidity assumption, and an internal alignment construct. The only explicit free settings are unstated hyperparameters; the paper reports no fitted physical parameters because it is a learning system rather than a derivation.

free parameters (1)

Joint optimization hyperparameters (photometric and semantic consistency loss weights, co-pruning thresholds) = Not reported in the supplied text
The joint objective in Eq. (2) and the collaborative co-pruning strategy are defined without specifying their hyperparameter values in the available text, so the method as presented contains undisclosed free settings.

assumptions (4)

domain assumption Known rigid per-object transformations and per-scan segmentation masks are available as input.
Section 3.2 states the goal uses scans (I_i, M_i) and 'known rigid object transformations' to align pairs; Eq. (5) applies per-object rigid transforms. No estimation or uncertainty handling is described.
domain assumption Objects transform rigidly between scans and the background remains static.
Eq. (5) defines T as per-object rigid transformations with background unchanged; non-rigid interaction would violate the model.
domain assumption Cross-scan appearance and segmentation are consistent enough for bi-directional photometric and semantic supervision.
Section 3.2 and ablation B assume that enforcing consistency between scans refines occluded regions; lighting or sensor changes across scans would break this.
standard math Alpha-blending rendering equations for color and segmentation (Eq. 1) are taken as the underlying scene model.
Borrowed from Gaussian Grouping (Ye et al. 2023); the paper does not re-derive them.

invented entities (1)

Pseudo-intermediate Gaussian state
purpose: A unified reference frame to align Gaussian fields captured under different object arrangements.
Introduced as an optimization construct (Section 3.3 and ablation P); it has no falsifiable handle outside the method's own renderings, and its benefit is shown only through the PSNR and SSIM ablation in Table 3.

how reviews work

0 comments

Cite this review

Pith. "Pith review of IGFuse: Interactive 3D Gaussian Scene Reconstruction via Multi-Scans Fusion." pith.science (2026). https://pith.science/paper/IPKZKD2L

@misc{pith2026250813153,
  author       = {Pith},
  title        = {Pith review of: IGFuse: Interactive 3D Gaussian Scene Reconstruction via Multi-Scans Fusion},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/IPKZKD2L}},
  note         = {Machine review of arXiv:2508.13153}
}

read the original abstract

Reconstructing complete and interactive 3D scenes remains a fundamental challenge in computer vision and robotics, particularly due to persistent object occlusions and limited sensor coverage. Multiview observations from a single scene scan often fail to capture the full structural details. Existing approaches typically rely on multi stage pipelines, such as segmentation, background completion, and inpainting or require per-object dense scanning, both of which are error-prone, and not easily scalable. We propose IGFuse, a novel framework that reconstructs interactive Gaussian scene by fusing observations from multiple scans, where natural object rearrangement between captures reveal previously occluded regions. Our method constructs segmentation aware Gaussian fields and enforces bi-directional photometric and semantic consistency across scans. To handle spatial misalignments, we introduce a pseudo-intermediate scene state for unified alignment, alongside collaborative co-pruning strategies to refine geometry. IGFuse enables high fidelity rendering and object level scene manipulation without dense observations or complex pipelines. Extensive experiments validate the framework's strong generalization to novel scene configurations, demonstrating its effectiveness for real world 3D reconstruction and real-to-simulation transfer. Our project page is available online.

Review history (2 revisions) →

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

41 extracted references · 10 canonical work pages

[1]

, " * write output.state after.block = add.period write newline

ENTRY address archivePrefix author booktitle chapter edition editor eid eprint howpublished institution isbn journal key month note number organization pages publisher school series title type volume year label extra.label sort.label short.list INTEGERS output.state before.all mid.sentence after.sentence after.block FUNCTION init.state.consts #0 'before.a...
[2]

write newline

" write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 global.max substring 't := if while FUNCTION word.in bbl.in capitalize " " * FUNCT...
[3]

Barcellona, L.; Zadaianchuk, A.; Allegro, D.; Papa, S.; Ghidoni, S.; and Gavves, E. 2024. Dream to Manipulate: Compositional World Models Empowering Robot Imitation Learning with Imagination. arXiv preprint arXiv:2412.14957

arXiv 2024
[4]

Blender Online Community . 2023. Blender - a 3D modelling and rendering package. https://www.blender.org

work page 2023
[5]

BlenderKit. 2023. BlenderKit: 3D Asset Library for Blender. https://www.blenderkit.com. Accessed in 2023

work page 2023
[6]

Cao, C.; Yu, C.; Wang, F.; Xue, X.; and Fu, Y. 2024 a . Mvinpainter: Learning multi-view consistent inpainting to bridge 2d and 3d editing. arXiv preprint arXiv:2408.08000

arXiv 2024
[7]

Cao, J.; Guan, S.; Ge, Y.; Li, W.; Yang, X.; and Ma, C. 2024 b . NeuMA: Neural material adaptor for visual grounding of intrinsic dynamics. Advances in Neural Information Processing Systems, 37: 65643--65669

work page 2024
[8]

Chen, Y.; Chen, Z.; Zhang, C.; Wang, F.; Yang, X.; Wang, Y.; Cai, Z.; Yang, L.; Liu, H.; and Lin, G. 2024. Gaussianeditor: Swift and controllable 3d editing with gaussian splatting. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 21476--21485

work page 2024

Show all 41 references

[9]

K.; Oh, S

Cheng, H. K.; Oh, S. W.; Price, B.; Schwing, A.; and Lee, J.-Y. 2023. Tracking anything with decoupled video segmentation. In Proceedings of the IEEE/CVF International Conference on Computer Vision, 1316--1326

2023
[10]

Han, X.; Liu, M.; Chen, Y.; Yu, J.; Lyu, X.; Tian, Y.; Wang, B.; Zhang, W.; and Pang, J. 2025. Re \^ 3 Sim: Generating High-Fidelity Simulation Data via 3D-Photorealistic Real-to-Sim for Robotic Manipulation. arXiv preprint arXiv:2502.08645

2025 arXiv
[11]

Hu, W.; Chai, W.; Hao, S.; Cui, X.; Wen, X.; Hwang, J.-N.; and Wang, G. 2025. Pointmap Association and Piecewise-Plane Constraint for Consistent and Compact 3D Gaussian Segmentation Field. arXiv preprint arXiv:2502.16303

2025 arXiv
[12]

Hu, X.; Wang, Y.; Fan, L.; Fan, J.; Peng, J.; Lei, Z.; Li, Q.; and Zhang, Z. 2024. Semantic anything in 3d gaussians. arXiv preprint arXiv:2401.17857

2024 arXiv
[13]

Huang, S.-Y.; Chou, Z.-T.; and Wang, Y.-C. F. 2025. 3D Gaussian Inpainting with Depth-Guided Cross-View Consistency. arXiv preprint arXiv:2502.11801

2025 arXiv
[14]

Jain, U.; Mirzaei, A.; and Gilitschenski, I. 2024. GaussianCut: Interactive segmentation via graph cut for 3D Gaussian Splatting. In The Thirty-eighth Annual Conference on Neural Information Processing Systems

2024
[15]

Kerbl, B.; Kopanas, G.; Leimk \"u hler, T.; and Drettakis, G. 2023. 3d gaussian splatting for real-time radiance field rendering. ACM Transactions on Graphics, 42(4): 1--14

2023
[16]

Li, H.; Wu, Y.; Meng, J.; Gao, Q.; Zhang, Z.; Wang, R.; and Zhang, J. 2024 a . InstanceGaussian: Appearance-Semantic Joint Gaussian Representation for 3D Instance-Level Perception. arXiv preprint arXiv:2411.19235

2024 arXiv
[17]

Li, X.; Li, J.; Zhang, Z.; Zhang, R.; Jia, F.; Wang, T.; Fan, H.; Tseng, K.-K.; and Wang, R. 2024 b . Robogsim: A real2sim2real robotic gaussian splatting simulator. arXiv preprint arXiv:2411.11839

2024 arXiv
[18]

L.; Xiao, J.; Zhu, K.; Xue, N.; Liu, Y.; Shen, Y.; and Cao, Y

Liu, Z.; Ouyang, H.; Wang, Q.; Cheng, K. L.; Xiao, J.; Zhu, K.; Xue, N.; Liu, Y.; Shen, Y.; and Cao, Y. 2024. Infusion: Inpainting 3d gaussians via learning depth completion from diffusion prior. arXiv preprint arXiv:2404.11613

2024 arXiv
[19]

Lou, H.; Liu, Y.; Pan, Y.; Geng, Y.; Chen, J.; Ma, W.; Li, C.; Wang, L.; Feng, H.; Shi, L.; et al. 2024. Robo-gs: A physics consistent spatial-temporal model for robotic arm with hybrid representation. arXiv preprint arXiv:2408.14873

2024 arXiv
[20]

Lyu, W.; Li, X.; Kundu, A.; Tsai, Y.-H.; and Yang, M.-H. 2024. Gaga: Group Any Gaussians via 3D-aware Memory Bank. arXiv preprint arXiv:2404.07977

2024 arXiv
[21]

Mendonca, R.; Bahl, S.; and Pathak, D. 2023. Structured world models from human videos. arXiv preprint arXiv:2308.10901

2023 arXiv
[22]

Oquab, M.; Darcet, T.; Moutakanni, T.; Vo, H.; Szafraniec, M.; Khalidov, V.; Fernandez, P.; Haziza, D.; Massa, F.; El-Nouby, A.; et al. 2023. Dinov2: Learning robust visual features without supervision. arXiv preprint arXiv:2304.07193

2023 arXiv
[23]

Pang, J.-C.; Tang, N.; Li, K.; Tang, Y.; Cai, X.-Q.; Zhang, Z.-Y.; Niu, G.; Sugiyama, M.; and Yu, Y. 2025. Learning View-invariant World Models for Visual Robotic Manipulation. In The Thirteenth International Conference on Learning Representations

2025
[24]

N.; Garg, S.; Yandun, F.; Held, D.; Kantor, G.; and Silwal, A

Qureshi, M. N.; Garg, S.; Yandun, F.; Held, D.; Kantor, G.; and Silwal, A. 2024. Splatsim: Zero-shot sim2real transfer of rgb manipulation policies using gaussian splatting. arXiv preprint arXiv:2409.10161

2024 arXiv
[25]

Ravi, N.; Gabeur, V.; Hu, Y.-T.; Hu, R.; Ryali, C.; Ma, T.; Khedr, H.; R \"a dle, R.; Rolland, C.; Gustafson, L.; et al. 2024. Sam 2: Segment anything in images and videos. arXiv preprint arXiv:2408.00714

2024 arXiv
[26]

Shen, Q.; Yang, X.; and Wang, X. 2024. Flashsplat: 2d to 3d gaussian splatting segmentation solved optimally. In European Conference on Computer Vision, 456--472. Springer

2024
[27]

Wang, G.; Pan, L.; Peng, S.; Liu, S.; Xu, C.; Miao, Y.; Zhan, W.; Tomizuka, M.; Pollefeys, M.; and Wang, H. 2024. Nerf in robotics: A survey. arXiv preprint arXiv:2405.01333

2024 arXiv
[28]

Wang, M.; Zhang, Y.; Ma, R.; Xu, W.; Zou, C.; and Morris, D. 2025. DecoupledGaussian: Object-Scene Decoupling for Physics-Based Interaction. arXiv preprint arXiv:2503.05484

2025 arXiv
[29]

Wu, J.; Yin, S.; Feng, N.; He, X.; Li, D.; Hao, J.; and Long, M. 2024 a . ivideogpt: Interactive videogpts are scalable world models. Advances in Neural Information Processing Systems, 37: 68082--68119

2024
[30]

Wu, Y.; Meng, J.; Li, H.; Wu, C.; Shi, Y.; Cheng, X.; Zhao, C.; Feng, H.; Ding, E.; Wang, J.; et al. 2024 b . OpenGaussian: Towards Point-Level 3D Gaussian-based Open Vocabulary Understanding. arXiv preprint arXiv:2406.02058

2024 arXiv
[31]

Xie, T.; Zong, Z.; Qiu, Y.; Li, X.; Feng, Y.; Yang, Y.; and Jiang, C. 2024. Physgaussian: Physics-integrated 3d gaussians for generative dynamics. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 4389--4398

2024
[32]

Yang, M.; Du, Y.; Ghasemipour, K.; Tompson, J.; Schuurmans, D.; and Abbeel, P. 2023. Learning interactive real-world simulators. arXiv preprint arXiv:2310.06114, 1(2): 6

2023 arXiv
[33]

Yang, S.; Yu, W.; Zeng, J.; Lv, J.; Ren, K.; Lu, C.; Lin, D.; and Pang, J. 2025. Novel Demonstration Generation with Gaussian Splatting Enables Robust One-Shot Manipulation. arXiv preprint arXiv:2504.13175

2025 arXiv
[34]

Ye, M.; Danelljan, M.; Yu, F.; and Ke, L. 2023. Gaussian grouping: Segment and edit anything in 3d scenes. arXiv preprint arXiv:2312.00732

2023 arXiv
[35]

A.; Cheng, R.; Irshad, M

Yu, J.; Fu, L.; Huang, H.; El-Refai, K.; Ambrus, R. A.; Cheng, R.; Irshad, M. Z.; and Goldberg, K. 2025. Real2Render2Real: Scaling Robot Data Without Dynamics Simulation or Robot Hardware. arXiv preprint arXiv:2505.09601

2025 arXiv
[36]

Zhang, J.; Jiang, J.; Chen, Y.; Jiang, K.; and Liu, X. 2025. COB-GS: Clear Object Boundaries in 3DGS Segmentation Based on Boundary-Adaptive Gaussian Splitting. arXiv preprint arXiv:2503.19443

2025 arXiv
[37]

Zhang, J.; Li, J.; Yu, X.; Huang, L.; Gu, L.; Zheng, J.; and Bai, X. 2024. Cor-gs: sparse-view 3d gaussian splatting via co-regularization. In European Conference on Computer Vision, 335--352. Springer

2024
[38]

Zhong, L.; Yu, H.-X.; Wu, J.; and Li, Y. 2024. Reconstruction and simulation of elastic objects with spring-mass 3d gaussians. In European Conference on Computer Vision, 407--423. Springer

2024
[39]

Zhu, R.; Qiu, S.; Liu, Z.; Hui, K.-H.; Wu, Q.; Heng, P.-A.; and Fu, C.-W. 2025 a . Rethinking End-to-End 2D to 3D Scene Segmentation in Gaussian Splatting. arXiv preprint arXiv:2503.14029

2025 arXiv
[40]

Zhu, S.; Mou, L.; Li, D.; Ye, B.; Huang, R.; and Zhao, H. 2025 b . VR-Robo: A Real-to-Sim-to-Real Framework for Visual Robot Navigation and Locomotion. arXiv preprint arXiv:2502.01536

2025 arXiv
[41]

Zhu, S.; Wang, G.; Kong, X.; Kong, D.; and Wang, H. 2024. 3d gaussian splatting in robotics: A survey. arXiv preprint arXiv:2410.12262

2024 arXiv

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.

[1] [1]

, " * write output.state after.block = add.period write newline

ENTRY address archivePrefix author booktitle chapter edition editor eid eprint howpublished institution isbn journal key month note number organization pages publisher school series title type volume year label extra.label sort.label short.list INTEGERS output.state before.all mid.sentence after.sentence after.block FUNCTION init.state.consts #0 'before.a...

[2] [2]

write newline

" write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 global.max substring 't := if while FUNCTION word.in bbl.in capitalize " " * FUNCT...

[3] [3]

Barcellona, L.; Zadaianchuk, A.; Allegro, D.; Papa, S.; Ghidoni, S.; and Gavves, E. 2024. Dream to Manipulate: Compositional World Models Empowering Robot Imitation Learning with Imagination. arXiv preprint arXiv:2412.14957

arXiv 2024

[4] [4]

Blender Online Community . 2023. Blender - a 3D modelling and rendering package. https://www.blender.org

work page 2023

[5] [5]

BlenderKit. 2023. BlenderKit: 3D Asset Library for Blender. https://www.blenderkit.com. Accessed in 2023

work page 2023

[6] [6]

Cao, C.; Yu, C.; Wang, F.; Xue, X.; and Fu, Y. 2024 a . Mvinpainter: Learning multi-view consistent inpainting to bridge 2d and 3d editing. arXiv preprint arXiv:2408.08000

arXiv 2024

[7] [7]

Cao, J.; Guan, S.; Ge, Y.; Li, W.; Yang, X.; and Ma, C. 2024 b . NeuMA: Neural material adaptor for visual grounding of intrinsic dynamics. Advances in Neural Information Processing Systems, 37: 65643--65669

work page 2024

[8] [8]

Chen, Y.; Chen, Z.; Zhang, C.; Wang, F.; Yang, X.; Wang, Y.; Cai, Z.; Yang, L.; Liu, H.; and Lin, G. 2024. Gaussianeditor: Swift and controllable 3d editing with gaussian splatting. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 21476--21485

work page 2024