pith. sign in

arxiv: 2506.17212 · v2 · submitted 2025-06-20 · 💻 cs.CV · cs.AI· cs.LG· cs.RO

Part²GS: Part-aware Modeling of Articulated Objects using 3D Gaussian Splatting

Tianjiao Yu , Vedant Shah , Muntasir Wahed , Ying Shen , Kiet A. Nguyen , Ismini Lourentzou This is my paper

Pith reviewed 2026-05-19 08:11 UTC · model grok-4.3

classification 💻 cs.CV cs.AIcs.LGcs.RO
keywords articulated objects3D Gaussian splattingpart-aware modelingphysics-based constraintsdigital twins3D reconstructionmotion consistency
0
0 comments X

The pith

Part²GS uses part-aware 3D Gaussians and physics constraints to model articulated objects with high-fidelity geometry and consistent motion.

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

The paper presents Part²GS as a framework for building detailed digital models of objects made of multiple moving parts. It starts from 3D Gaussian splatting but adds explicit part awareness so each component can transform independently while keeping sharp geometry. Physics rules are added to guide the motion, including rules that keep parts in contact, match their speeds, align their directions, and push them apart to avoid overlaps. A reader would care because this combination aims to produce models that look realistic both when still and when moving, which matters for creating usable digital copies of real machines or furniture without requiring separate motion recordings for every new object.

Core claim

Part²GS leverages a part-aware 3D Gaussian representation that encodes articulated components with learnable attributes, enabling structured, disentangled transformations that preserve high-fidelity geometry. To ensure physically consistent motion, it proposes a motion-aware canonical representation guided by physics-based constraints, including contact enforcement, velocity consistency, and vector-field alignment, together with a field of repel points to prevent part collisions and maintain stable articulation paths.

What carries the argument

A part-aware 3D Gaussian representation that assigns learnable attributes to individual components, paired with physics-based motion constraints and a repel point field.

If this is right

  • Movable parts achieve up to ten times lower Chamfer Distance error than prior methods on both synthetic and real data.
  • Geometry stays sharp and detailed even while parts articulate.
  • Motion stays coherent because contact, velocity, and direction rules are enforced together.
  • Collisions between parts are reduced by the repel point field.
  • The same model works without change on both computer-generated and camera-captured scenes.

Where Pith is reading between the lines

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

  • If the constraints prove robust, the same physics rules could be transferred to other point-based or mesh-based representations of moving objects.
  • The repel point idea might extend naturally to modeling joint limits or friction in more complex machines.
  • Successful part disentanglement could simplify downstream tasks such as editing individual components in a reconstructed scene.

Load-bearing premise

The physics-based constraints together with the repel point field are enough on their own to produce physically consistent articulation without extra real-world motion data or separate validation of those constraints.

What would settle it

Running the trained model on previously unseen motion sequences of real articulated objects and checking whether parts penetrate each other, violate velocity rules, or lose contact during the motion.

Figures

Figures reproduced from arXiv: 2506.17212 by Ismini Lourentzou, Kiet A. Nguyen, Muntasir Wahed, Tianjiao Yu, Vedant Shah, Ying Shen.

Figure 1
Figure 1. Figure 1: Overview of Part2GS. Given two sets of multi-view images of an object, we first reconstruct independent coarse 3D Gaussian models and learn a motion-informed, part-aware canonical Gaussian. We optimize the canonical Gaussian under physical constraints ℒphys and part awareness ℒpart. Finally, we learn the articulation model with repel points ℱrepel and ℒarticulation (details in §4). 4. Part2GS: Part-aware O… view at source ↗
Figure 2
Figure 2. Figure 2: Physical Constraints. (1) Contact Loss penalizes interpenetration by minimizing the angle between two vectors for each Gaussian: a) the vector pointing to the center of the opposing part, and b) the vector pointing to its nearest Gaussian in that part. Red dots (•) denote object centers. (2) Velocity Consistency encourages coherent motion trajectories (e.g., µ 0 i == µ 1 i ). Red dots (•) represent the sam… view at source ↗
Figure 3
Figure 3. Figure 3: Qualitative Comparison of Part Discovery Across Object States (columns) and Discovery [PITH_FULL_IMAGE:figures/full_fig_p011_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Qualitative Results on objects with different joints and distinct geometry structures. undergoing continuous motion, with smooth transitions between configurations. These intermediate frames demonstrate that Part2GS produces consistent motion paths through the full articulation sequence [PITH_FULL_IMAGE:figures/full_fig_p018_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Qualitative Results [PITH_FULL_IMAGE:figures/full_fig_p019_5.png] view at source ↗
read the original abstract

Articulated objects are common in the real world, yet modeling their structure and motion remains a challenging task for 3D reconstruction methods. In this work, we introduce Part$^{2}$GS, a novel framework for modeling articulated digital twins of multi-part objects with high-fidelity geometry and physically consistent articulation. Part$^{2}$GS leverages a part-aware 3D Gaussian representation that encodes articulated components with learnable attributes, enabling structured, disentangled transformations that preserve high-fidelity geometry. To ensure physically consistent motion, we propose a motion-aware canonical representation guided by physics-based constraints, including contact enforcement, velocity consistency, and vector-field alignment. Furthermore, we introduce a field of repel points to prevent part collisions and maintain stable articulation paths, significantly improving motion coherence over baselines. Extensive evaluations on both synthetic and real-world datasets show that Part$^{2}$GS consistently outperforms state-of-the-art methods by up to 10$\times$ in Chamfer Distance for movable parts.

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

3 major / 2 minor

Summary. The manuscript introduces Part²GS, a framework for part-aware modeling of articulated objects via 3D Gaussian Splatting. It encodes components with learnable attributes to enable disentangled transformations that preserve geometry, and incorporates physics-based constraints (contact enforcement, velocity consistency, vector-field alignment) plus a repel point field to promote physically consistent articulation without explicit real-world motion supervision. Extensive evaluations on synthetic and real-world datasets are reported to yield up to 10× lower Chamfer Distance on movable parts relative to prior state-of-the-art methods.

Significance. If the central results hold under rigorous validation, the work would advance 3D reconstruction of articulated objects by integrating high-fidelity Gaussian representations with explicit physics guidance for motion coherence. This direction is timely given the adoption of 3D Gaussians in dynamic scene modeling and could support applications in robotics and simulation that require stable part interactions.

major comments (3)
  1. [Abstract, §4] Abstract and §4 (Experiments): The claim of consistent outperformance 'by up to 10× in Chamfer Distance for movable parts' lacks accompanying error bars, precise dataset statistics, protocol details, or per-baseline breakdowns. Without these, the magnitude and robustness of the reported gains cannot be assessed from the provided evaluation description.
  2. [§3.2, §4.3] §3.2 (Physics Constraints) and §4.3 (Ablations): No ablation isolating the contribution of contact enforcement, velocity consistency, vector-field alignment, or the repel point field is described. Because Chamfer Distance measures surface geometry rather than physical properties (e.g., inter-part penetration volume or adherence to non-penetration), it remains unclear whether the physics terms drive the claimed articulation consistency or whether gains arise primarily from the part-aware disentanglement.
  3. [§3.3] §3.3 (Repel Point Field): The repel point field is introduced to prevent collisions, yet the manuscript provides no quantitative validation (e.g., penetration metrics or trajectory stability measures) demonstrating that this component produces physically consistent paths beyond what the canonical representation alone achieves.
minor comments (2)
  1. [Abstract] The abstract uses 'Part$^{2}$GS' without an immediate expansion or reference to the full name on first use; this should be clarified in the introduction for readability.
  2. [§3] Notation for learnable attributes per part and the vector field should be defined consistently between the method section and equations to avoid ambiguity.

Simulated Author's Rebuttal

3 responses · 0 unresolved

We thank the referee for the constructive and detailed feedback. We address each major comment point by point below, indicating planned revisions to strengthen the manuscript.

read point-by-point responses

  1. Referee: [Abstract, §4] Abstract and §4 (Experiments): The claim of consistent outperformance 'by up to 10× in Chamfer Distance for movable parts' lacks accompanying error bars, precise dataset statistics, protocol details, or per-baseline breakdowns. Without these, the magnitude and robustness of the reported gains cannot be assessed from the provided evaluation description.

    Authors: We agree that additional statistical details are needed for full assessment of the results. In the revised manuscript, we will add error bars (standard deviations over multiple runs), precise dataset statistics (object counts, sequence lengths, train/test splits), a clear evaluation protocol description, and per-baseline Chamfer Distance breakdowns for movable parts in both the abstract and Section 4. revision: yes

  2. Referee: [§3.2, §4.3] §3.2 (Physics Constraints) and §4.3 (Ablations): No ablation isolating the contribution of contact enforcement, velocity consistency, vector-field alignment, or the repel point field is described. Because Chamfer Distance measures surface geometry rather than physical properties (e.g., inter-part penetration volume or adherence to non-penetration), it remains unclear whether the physics terms drive the claimed articulation consistency or whether gains arise primarily from the part-aware disentanglement.

    Authors: We acknowledge that finer-grained ablations would clarify individual contributions. While §4.3 reports ablations on the combined physics constraints, we will add new experiments isolating contact enforcement, velocity consistency, vector-field alignment, and the repel point field. We will also report direct physical metrics such as inter-part penetration volume alongside Chamfer Distance to better demonstrate the physics terms' role in articulation consistency beyond geometric improvements from part disentanglement. revision: yes

  3. Referee: [§3.3] §3.3 (Repel Point Field): The repel point field is introduced to prevent collisions, yet the manuscript provides no quantitative validation (e.g., penetration metrics or trajectory stability measures) demonstrating that this component produces physically consistent paths beyond what the canonical representation alone achieves.

    Authors: We thank the referee for highlighting this gap. In the revised §4.3, we will include quantitative validation using penetration volume metrics and trajectory stability measures (e.g., variance in inter-part distances and velocities over time). These will be reported for the full model versus an ablation without the repel point field to show its specific contribution to collision-free, stable articulation paths. revision: yes

Circularity Check

0 steps flagged

No circularity: derivation relies on proposed representations and external physics constraints

full rationale

The paper presents Part²GS as a framework combining a part-aware 3D Gaussian representation with learnable attributes and physics-based constraints (contact enforcement, velocity consistency, vector-field alignment, repel point field) to achieve physically consistent articulation. These elements are introduced as novel contributions guided by established physical principles rather than being derived from or fitted to the target outputs. No equations or steps in the provided abstract reduce predictions or results to inputs by construction, self-definition, or self-citation chains. Evaluations on synthetic and real datasets use independent geometric metrics (Chamfer Distance), and the central claims remain independent of any load-bearing self-referential loops. This is a standard non-circular finding for a methods paper introducing new modeling components.

Axiom & Free-Parameter Ledger

1 free parameters · 1 axioms · 1 invented entities

The central claim rests on the effectiveness of learnable part attributes and newly introduced physics constraints and repel points; these are not derived from first principles but postulated to enforce consistency.

free parameters (1)
  • learnable attributes per part
    Encoded in the part-aware 3D Gaussian representation to enable disentangled transformations.
axioms (1)
  • domain assumption Physics-based constraints (contact enforcement, velocity consistency, vector-field alignment) produce physically consistent motion
    Invoked to guide the motion-aware canonical representation.
invented entities (1)
  • field of repel points no independent evidence
    purpose: Prevent part collisions and maintain stable articulation paths
    New mechanism introduced to improve motion coherence over baselines.

pith-pipeline@v0.9.0 · 5734 in / 1300 out tokens · 35941 ms · 2026-05-19T08:11:22.619692+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

52 extracted references · 52 canonical work pages · 2 internal anchors

  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]

    Procthor: Large-scale embodied ai using procedural generation.Advances in Neural Information Processing Systems (NeurIPS), 2022

    Matt Deitke, Eli VanderBilt, Alvaro Herrasti, Luca Weihs, Kiana Ehsani, Jordi Salvador, Winson Han, Eric Kolve, Aniruddha Kembhavi, and Roozbeh Mottaghi. Procthor: Large-scale embodied ai using procedural generation.Advances in Neural Information Processing Systems (NeurIPS), 2022

    work page 2022

  3. [3]

    Objaverse: A universe of annotated 3d objects

    MattDeitke, DustinSchwenk, JordiSalvador, LucaWeihs, OscarMichel, EliVanderBilt, LudwigSchmidt, Kiana Ehsani, Aniruddha Kembhavi, and Ali Farhadi. Objaverse: A universe of annotated 3d objects. In IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2023

    work page 2023

  4. [4]

    Banana: Banach fixed-point network for pointcloud segmentation with inter-part equivariance.Advances in Neural Information Processing Systems (NeurIPS), 2023

    Congyue Deng, Jiahui Lei, William B Shen, Kostas Daniilidis, and Leonidas J Guibas. Banana: Banach fixed-point network for pointcloud segmentation with inter-part equivariance.Advances in Neural Information Processing Systems (NeurIPS), 2023

    work page 2023

  5. [5]

    Act the part: Learning interaction strategies for articulated object part discovery

    Samir Yitzhak Gadre, Kiana Ehsani, and Shuran Song. Act the part: Learning interaction strategies for articulated object part discovery. InInternational Conference on Computer Vision (ICCV), 2021

    work page 2021

  6. [6]

    Gapart- net: Cross-category domain-generalizable object perception and manipulation via generalizable and actionable parts

    Haoran Geng, Helin Xu, Chengyang Zhao, Chao Xu, Li Yi, Siyuan Huang, and He Wang. Gapart- net: Cross-category domain-generalizable object perception and manipulation via generalizable and actionable parts. InIEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2023

    work page 2023

  7. [7]

    Articulatedgs: Self-supervised digital twin modeling of articulated objects using 3d gaussian splatting

    JunfuGuo, YuXin, GaoyiLiu, Kai Xu, LigangLiu, andRuizhenHu. Articulatedgs: Self-superviseddigital twin modeling of articulated objects using 3d gaussian splatting.arXiv preprint arXiv:2503.08135, 2025

    work page arXiv 2025

  8. [8]

    Carto: Category and joint agnostic reconstruction of articulated objects

    Nick Heppert, Muhammad Zubair Irshad, Sergey Zakharov, Katherine Liu, Rares Andrei Ambrus, Jeannette Bohg, Abhinav Valada, and Thomas Kollar. Carto: Category and joint agnostic reconstruction of articulated objects. InIEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2023

    work page 2023

  9. [9]

    Learning to predict part mobility from a single static snapshot.ACM Transactions on Graphics (TOG), 2017

    Ruizhen Hu, Wenchao Li, Oliver Van Kaick, Ariel Shamir, Hao Zhang, and Hui Huang. Learning to predict part mobility from a single static snapshot.ACM Transactions on Graphics (TOG), 2017

    work page 2017

  10. [10]

    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. InIEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2024

    work page 2024

  11. [11]

    Screwnet: Category-independent articulation model estimation from depth images using screw theory

    Ajinkya Jain, Rudolf Lioutikov, Caleb Chuck, and Scott Niekum. Screwnet: Category-independent articulation model estimation from depth images using screw theory. InInternational Conference on Robotics and Automation (ICRA), 2021

    work page 2021

  12. [12]

    Ditto: Building digital twins of articulated objects from interaction

    Zhenyu Jiang, Cheng-Chun Hsu, and Yuke Zhu. Ditto: Building digital twins of articulated objects from interaction. InIEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2022. 12

    work page 2022

  13. [13]

    Deformable 3d gaussian splatting for animatable human avatars.Computing Research Repository, 2023

    HyunJun Jung, Nikolas Brasch, Jifei Song, Eduardo Pérez-Pellitero, Yiren Zhou, Zhihao Li, Nassir Navab, and Benjamin Busam. Deformable 3d gaussian splatting for animatable human avatars.Computing Research Repository, 2023

    work page 2023

  14. [14]

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

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

    work page 2023

  15. [15]

    AI2-THOR: An Interactive 3D Environment for Visual AI

    Eric Kolve, Roozbeh Mottaghi, Winson Han, Eli VanderBilt, Luca Weihs, Alvaro Herrasti, Matt Deitke, Kiana Ehsani, Daniel Gordon, Yuke Zhu, et al. Ai2-thor: An interactive 3d environment for visual ai. arXiv preprint arXiv:1712.05474, 2017

    work page internal anchor Pith review Pith/arXiv arXiv 2017

  16. [16]

    Articulate-anything: Automatic modeling of articulated objects via a vision-language foundation model

    Long Le, Jason Xie, William Liang, Hung-Ju Wang, Yue Yang, Yecheng Jason Ma, Kyle Vedder, Arjun Krishna, Dinesh Jayaraman, and Eric Eaton. Articulate-anything: Automatic modeling of articulated objects via a vision-language foundation model. InInternational Conference on Learning Representations (ICLR), 2025

    work page 2025

  17. [17]

    Nap: Neural 3d articulated object prior.Advances in Neural Information Processing Systems (NeurIPS), 2023

    Jiahui Lei, Congyue Deng, William B Shen, Leonidas J Guibas, and Kostas Daniilidis. Nap: Neural 3d articulated object prior.Advances in Neural Information Processing Systems (NeurIPS), 2023

    work page 2023

  18. [18]

    Mobility fitting using 4d ransac

    Hao Li, Guowei Wan, Honghua Li, Andrei Sharf, Kai Xu, and Baoquan Chen. Mobility fitting using 4d ransac. InComputer Graphics Forum, 2016

    work page 2016

  19. [19]

    Garf: Learning generalizable 3d reassembly for real-world fractures

    Sihang Li, Zeyu Jiang, Grace Chen, Chenyang Xu, Siqi Tan, Xue Wang, Irving Fang, Kristof Zyskowski, Shannon P McPherron, Radu Iovita, et al. Garf: Learning generalizable 3d reassembly for real-world fractures. arXiv preprint arXiv:2504.05400, 2025

    work page arXiv 2025

  20. [20]

    Category-level articulated object pose estimation

    Xiaolong Li, He Wang, Li Yi, Leonidas J Guibas, A Lynn Abbott, and Shuran Song. Category-level articulated object pose estimation. InIEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2020

    work page 2020

  21. [21]

    Spacetime gaussian feature splatting for real-time dynamic view synthesis

    Zhan Li, Zhang Chen, Zhong Li, and Yi Xu. Spacetime gaussian feature splatting for real-time dynamic view synthesis. InIEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2024

    work page 2024

  22. [22]

    Flow Matching Guide and Code

    Yaron Lipman, Marton Havasi, Peter Holderrieth, Neta Shaul, Matt Le, Brian Karrer, Ricky TQ Chen, David Lopez-Paz, Heli Ben-Hamu, and Itai Gat. Flow matching guide and code. arXiv preprint arXiv:2412.06264, 2024

    work page internal anchor Pith review Pith/arXiv arXiv 2024

  23. [23]

    Semi-weakly supervised object kinematic motion prediction

    Gengxin Liu, Qian Sun, Haibin Huang, Chongyang Ma, Yulan Guo, Li Yi, Hui Huang, and Ruizhen Hu. Semi-weakly supervised object kinematic motion prediction. InIEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2023

    work page 2023

  24. [24]

    Paris: Part-level reconstruction and motion analysis for articulated objects

    Jiayi Liu, Ali Mahdavi-Amiri, and Manolis Savva. Paris: Part-level reconstruction and motion analysis for articulated objects. InInternational Conference on Computer Vision (ICCV), 2023

    work page 2023

  25. [25]

    Cage: controllable articulation generation

    Jiayi Liu, Hou In Ivan Tam, Ali Mahdavi-Amiri, and Manolis Savva. Cage: controllable articulation generation. InIEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2024

    work page 2024

  26. [26]

    SINGAPO: Single image controlled generation of articulated parts in objects

    Jiayi Liu, Denys Iliash, Angel X Chang, Manolis Savva, and Ali Mahdavi Amiri. SINGAPO: Single image controlled generation of articulated parts in objects. InInternational Conference on Learning Representations (ICLR), 2025. 13

    work page 2025

  27. [27]

    AKB-48: A real-world articulated object knowledge base

    Liu Liu, Wenqiang Xu, Haoyuan Fu, Sucheng Qian, Qiaojun Yu, Yang Han, and Cewu Lu. AKB-48: A real-world articulated object knowledge base. InIEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2022

    work page 2022

  28. [28]

    Flow straight and fast: Learning to generate and transfer data with rectified flow

    Xingchao Liu, Chengyue Gong, and qiang liu. Flow straight and fast: Learning to generate and transfer data with rectified flow. InInternational Conference on Learning Representations (ICLR), 2023

    work page 2023

  29. [29]

    Self-supervised category-level articulated object pose estimation with part-level SE(3) equivariance

    Xueyi Liu, Ji Zhang, Ruizhen Hu, Haibin Huang, He Wang, and Li Yi. Self-supervised category-level articulated object pose estimation with part-level SE(3) equivariance. InInternational Conference on Learning Representations (ICLR), 2023

    work page 2023

  30. [30]

    Building interactable replicas of complex articulated objects via gaussian splatting

    Yu Liu, Baoxiong Jia, Ruijie Lu, Junfeng Ni, Song-Chun Zhu, and Siyuan Huang. Building interactable replicas of complex articulated objects via gaussian splatting. InInternational Conference on Learning Representations (ICLR), 2025

    work page 2025

  31. [31]

    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. InIEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2024

    work page 2024

  32. [32]

    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. InInternational Conference on 3D Vision, 2024

    work page 2024

  33. [33]

    Multiscan: Scalable rgbd scanning for 3d environments with articulated objects.Advances in Neural Information Processing Systems (NeurIPS), 2022

    Yongsen Mao, Yiming Zhang, Hanxiao Jiang, Angel Chang, and Manolis Savva. Multiscan: Scalable rgbd scanning for 3d environments with articulated objects.Advances in Neural Information Processing Systems (NeurIPS), 2022

    work page 2022

  34. [34]

    Nerf: Representing scenes as neural radiance fields for view synthesis.Communications of the ACM, 2021

    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.Communications of the ACM, 2021

    work page 2021

  35. [35]

    Illustrating how mechanical assemblies work.ACM Transactions on Graphics (TOG), 2010

    Niloy J Mitra, Yong-Liang Yang, Dong-Ming Yan, Wilmot Li, Maneesh Agrawala, et al. Illustrating how mechanical assemblies work.ACM Transactions on Graphics (TOG), 2010

    work page 2010

  36. [36]

    Where2act: From pixels to actions for articulated 3D objects

    Kaichun Mo, Leonidas J Guibas, Mustafa Mukadam, Abhinav Gupta, and Shubham Tulsiani. Where2act: From pixels to actions for articulated 3D objects. InInternational Conference on Computer Vision (ICCV), 2021

    work page 2021

  37. [37]

    Habitat 3.0: A co-habitat for humans, avatars, and robots

    Xavier Puig, Eric Undersander, Andrew Szot, Mikael Dallaire Cote, Tsung-Yen Yang, Ruslan Partsey, Ruta Desai, Alexander Clegg, Michal Hlavac, So Yeon Min, Vladimír Vondruš, Theophile Gervet, Vincent- Pierre Berges, John M Turner, Oleksandr Maksymets, Zsolt Kira, Mrinal Kalakrishnan, Jitendra Malik, Devendra Singh Chaplot, Unnat Jain, Dhruv Batra, Akshara ...

    work page 2024

  38. [38]

    Understanding 3d object interaction from a single image

    Shengyi Qian and David F Fouhey. Understanding 3d object interaction from a single image. In International Conference on Computer Vision (ICCV), 2023

    work page 2023

  39. [39]

    3dgs-avatar: Animatable avatars via deformable 3d gaussian splatting

    Zhiyin Qian, Shaofei Wang, Marko Mihajlovic, Andreas Geiger, and Siyu Tang. 3dgs-avatar: Animatable avatars via deformable 3d gaussian splatting. InIEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2024. 14

    work page 2024

  40. [40]

    Mobility-trees for indoor scenes manipulation

    Andrei Sharf, Hui Huang, Cheng Liang, Jiapei Zhang, Baoquan Chen, and Minglun Gong. Mobility-trees for indoor scenes manipulation. InComputer Graphics Forum, 2014

    work page 2014

  41. [41]

    Self-supervised learning of part mobility from point cloud sequence

    Yahao Shi, Xinyu Cao, and Bin Zhou. Self-supervised learning of part mobility from point cloud sequence. InComputer Graphics Forum, 2021

    work page 2021

  42. [42]

    Reacto: Reconstructing articulated objects from a single video

    Chaoyue Song, Jiacheng Wei, Chuan Sheng Foo, Guosheng Lin, and Fayao Liu. Reacto: Reconstructing articulated objects from a single video. InIEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2024

    work page 2024

  43. [43]

    Leia: Latent view-invariant embeddings for implicit 3d articulation

    Archana Swaminathan, Anubhav Gupta, Kamal Gupta, Shishira R Maiya, Vatsal Agarwal, and Abhinav Shrivastava. Leia: Latent view-invariant embeddings for implicit 3d articulation. InEuropean Conference on Computer Vision (ECCV), 2024

    work page 2024

  44. [44]

    Cg3d: Compositional generation for text-to-3d via gaussian splatting.arXiv preprint arXiv:2311.17907, 2023

    Alexander Vilesov, Pradyumna Chari, and Achuta Kadambi. Cg3d: Compositional generation for text-to-3d via gaussian splatting.arXiv preprint arXiv:2311.17907, 2023

    work page arXiv 2023

  45. [45]

    Superpoint gaussian splatting for real-time high-fidelity dynamic scene reconstruction

    Diwen Wan, Ruijie Lu, and Gang Zeng. Superpoint gaussian splatting for real-time high-fidelity dynamic scene reconstruction. InInternational Conference on Machine Learning (ICML), 2024

    work page 2024

  46. [46]

    Template-free articulated gaussian splatting for real-time reposable dynamic view synthesis

    Diwen Wan, Yuxiang Wang, Ruijie Lu, and Gang Zeng. Template-free articulated gaussian splatting for real-time reposable dynamic view synthesis. InAdvances in Neural Information Processing Systems (NeurIPS), 2024

    work page 2024

  47. [47]

    Neural implicit representation for building digital twins of unknown articulated objects

    Yijia Weng, Bowen Wen, Jonathan Tremblay, Valts Blukis, Dieter Fox, Leonidas Guibas, and Stan Birchfield. Neural implicit representation for building digital twins of unknown articulated objects. In IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2024

    work page 2024

  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. InIEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2024

    work page 2024

  49. [49]

    Chang, Leonidas J

    Fanbo Xiang, Yuzhe Qin, Kaichun Mo, Yikuan Xia, Hao Zhu, Fangchen Liu, Minghua Liu, Hanxiao Jiang, Yifu Yuan, He Wang, Li Yi, Angel X. Chang, Leonidas J. Guibas, and Hao Su. SAPIEN: A simulated part-based interactive environment. InIEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2020

    work page 2020

  50. [50]

    Joint-aware manipulation of deformable models.ACM Transactions on Graphics (TOG), 2009

    Weiwei Xu, Jun Wang, KangKang Yin, Kun Zhou, Michiel Van De Panne, Falai Chen, and Baining Guo. Joint-aware manipulation of deformable models.ACM Transactions on Graphics (TOG), 2009

    work page 2009

  51. [51]

    Differentiable surface splatting for point-based geometry processing.ACM Transactions on Graphics (TOG), 2019

    Wang Yifan, Felice Serena, Shihao Wu, Cengiz Öztireli, and Olga Sorkine-Hornung. Differentiable surface splatting for point-based geometry processing.ACM Transactions on Graphics (TOG), 2019. 15 A. Experimental Details Datasets. We conduct evaluations across three distinct datasets, each designed to capture varying levels of articulation complexity: (i) P...

    work page 2019

  52. [52]

    system. (ii) DTA-Multi[47], a dataset that offers a moderate challenge with 2 synthetic objects from PartNet-Mobility, where each object contains a static component and two independently movable parts. (iii) ArtGS-Multi[30], a recent dataset that targets more intricate structures, featuring 5 synthetic articulated objects fromPartNet-Mobility, each compos...

    work page