pith. sign in

arxiv: 2606.18243 · v1 · pith:3A74LE76new · submitted 2026-06-16 · 💻 cs.CV · cs.GR· cs.RO

MOCHI: Motion Enhancement of Collaborative Human-object Interactions

Jiye Lee , Yonghun Choi , Jungdam Won This is my paper

Pith reviewed 2026-06-27 01:19 UTC · model grok-4.3

classification 💻 cs.CV cs.GRcs.RO
keywords motion enhancementhuman-object interactioncollaborative interactionsdiffusion modelsmotion optimizationgrasp generationMHOI data
0
0 comments X

The pith

MOCHI enhances noisy collaborative human-object interaction data through grasp optimization and diffusion-based motion refinement using single-person priors.

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

The paper introduces MOCHI, a two-stage framework designed to clean up artifacts in MHOI captures such as hand-object misalignments and motion jitter. In the first stage, it optimizes physically plausible hand grasps from noisy body poses and extends them to full sequences. The second stage refines the full-body motions of all participants via a diffusion model that incorporates interaction constraints into single-person motion priors. This matters because high-quality motion data is essential for training models that simulate realistic collaborative scenarios involving multiple people and shared objects. The method demonstrates robustness to different data sources and participant numbers while enabling new applications like keyframe editing.

Core claim

MOCHI is a two-stage pipeline that first generates physically plausible hand grasps through optimization from noisy body input and extends them into complete hand-object sequences, then refines full-body motions for all participants using a diffusion-based noise optimization framework augmented with objectives that encode human-object and human-human interaction information within single-person motion priors.

What carries the argument

Diffusion-based noise optimization framework that encodes human-object and human-human interactions into single-person motion priors.

If this is right

  • Works on data from existing capture methods or generative models.
  • Robust across varying numbers of participants and interaction types.
  • Supports applications such as keyframe-based MHOI creation.
  • Enables data augmentation by varying object geometries.

Where Pith is reading between the lines

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

  • If single-person priors can be augmented this way, similar techniques might apply to other multi-agent motion problems like team sports.
  • The approach could reduce reliance on specialized multi-person capture equipment.
  • Extending the method to real-time applications might improve interactive simulations in VR.

Load-bearing premise

Single-person motion priors can be augmented with additional objectives to encode the mutual anticipation and adjustments in collaborative interactions without introducing breaking artifacts.

What would settle it

Observing whether the optimized motions maintain consistent contacts and smooth trajectories in long sequences of complex multi-person object manipulations that were not used in training.

Figures

Figures reproduced from arXiv: 2606.18243 by Jiye Lee, Jungdam Won, Yonghun Choi.

Figure 1
Figure 1. Figure 1: Our method enhances noisy multi-human object interaction (MHOI) data ( [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: System Overview. human object interactions, detailed hand motions are often missing due to the inherent complexity of finger movements and the techni￾cal challenges of hand motion capture. Second, full-body motions frequently exhibit unnatural artifacts, primarily due to difficulties in capturing MHOI scenarios such as occlusions. Third, the complexity and diversity of real world MHOI scenarios far exceed … view at source ↗
Figure 3
Figure 3. Figure 3: An overview of grasping hand pose generation. (Left) Lower arms in [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: An example of a cylindrical bound for grabbing a table. (Left) The [PITH_FULL_IMAGE:figures/full_fig_p006_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Visualization of the interaction graph showing self-edges [PITH_FULL_IMAGE:figures/full_fig_p008_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Qualitative comparison: CORE4D-Original and Ours [PITH_FULL_IMAGE:figures/full_fig_p010_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: Qualitative comparison: CORE4D-Noisy and Ours. to the data, showing not only increased foot skating and jitter but also temporally inconsistent and physically implausible contacts caused by the added noise. CORE4D-HMR exhibits similar levels of artifacts as CORE4D-Noisy. Even for these more challenging capture setups, our method consistently improves both contact plausibility and motion naturalness. Visual… view at source ↗
Figure 8
Figure 8. Figure 8: The original MHOI data exhibit several artifacts, including [PITH_FULL_IMAGE:figures/full_fig_p010_8.png] view at source ↗
Figure 9
Figure 9. Figure 9: Qualitative comparison between CORE4D data with hand motion [PITH_FULL_IMAGE:figures/full_fig_p011_9.png] view at source ↗
Figure 11
Figure 11. Figure 11: Quantitative results (penetration metric) of augmenting chair [PITH_FULL_IMAGE:figures/full_fig_p012_11.png] view at source ↗
Figure 10
Figure 10. Figure 10: Qualitative results of enhancing OMOMO model’s MHOI output. [PITH_FULL_IMAGE:figures/full_fig_p012_10.png] view at source ↗
Figure 12
Figure 12. Figure 12: Qualitative results of augmenting chair-related MHOI sequences (rotate, hold, handover) into new chair objects in the ShapeNet dataset. [PITH_FULL_IMAGE:figures/full_fig_p013_12.png] view at source ↗
Figure 13
Figure 13. Figure 13: (Top) Initial motion constructed from keyframe interpolation. (Bottom) Enhanced MHOI motion generated by our pipeline. Frames with black boxes [PITH_FULL_IMAGE:figures/full_fig_p013_13.png] view at source ↗
Figure 14
Figure 14. Figure 14: (Top) Initial motion constructed from keyframe interpolation. (Bottom) Enhanced MHOI motion generated by our pipeline. Frames with black boxes [PITH_FULL_IMAGE:figures/full_fig_p014_14.png] view at source ↗
Figure 16
Figure 16. Figure 16: Qualitative results with MHOI motion where static object interac [PITH_FULL_IMAGE:figures/full_fig_p014_16.png] view at source ↗
Figure 18
Figure 18. Figure 18: Qualitative comparison between our method (middle) and random [PITH_FULL_IMAGE:figures/full_fig_p015_18.png] view at source ↗
Figure 19
Figure 19. Figure 19: Qualitative ablation study of diffusion optimization objectives. From [PITH_FULL_IMAGE:figures/full_fig_p016_19.png] view at source ↗
read the original abstract

Collaborative human-object interaction shows dynamic and complex movements that require mutual anticipation and continuous adjustment between participants and the shared object. Modeling such collaborative multi-human object interaction (MHOI) scenarios requires high-quality data acquisition as a foundational step; however, this is challenging due to the inherent complexity of MHOI where human-human and human-object interactions occur simultaneously. Such complexity leads to noisy MHOI captures characterized by several artifacts: contact misalignment between hands and objects, motion jitter and temporal inconsistencies in the captured sequences, and missing or incomplete finger-level articulation details. To address these challenges, we present MOCHI (MOtion Enhancement of Collaborative Human-object Interactions), a two-stage framework for enhancing noisy MHOI data. Our approach first generates physically plausible hand grasps through optimization from noisy body input, producing grasps that are both physically plausible and semantically consistent with the body pose, where these optimized grasps are extended into complete hand-object interaction sequences. Consequently, the full-body motion for all participants are refined through a diffusion-based noise optimization framework that uses single-person motion priors. During the optimization process, we introduce optimization objectives to encode human-object and human-human interaction information within these single-person priors. Experimental results demonstrate the effectiveness of our pipeline across diverse MHOI data, either acquired by existing capture methods or synthesized by generative models. We further show robustness of our system across varying numbers of participants and types of interactions, and demonstrate various applications including keyframe-based MHOI creation and data augmentation through varying object geometries.

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 paper introduces MOCHI, a two-stage framework for enhancing noisy collaborative multi-human object interaction (MHOI) captures. Stage 1 optimizes physically plausible and semantically consistent hand grasps from noisy body inputs and extends them into full hand-object sequences. Stage 2 refines full-body motions for all participants via diffusion-based noise optimization that augments single-person motion priors with additional objectives encoding human-object and human-human interaction information. The authors claim the pipeline is effective on data from existing capture methods or generative models, robust across varying participant counts and interaction types, and enables applications such as keyframe-based MHOI creation and data augmentation via object geometry variation.

Significance. If the central claims hold with rigorous quantitative support, the work would be significant for computer vision and graphics by providing a practical method to improve the quality of MHOI datasets, a known bottleneck for modeling complex collaborative dynamics. The combination of grasp optimization with diffusion priors augmented by interaction objectives represents a targeted approach to artifact removal while attempting to preserve multi-agent coordination; successful validation could directly benefit downstream tasks in animation, robotics, and interaction synthesis.

major comments (3)
  1. [Abstract] Abstract: the central claims of effectiveness and robustness across participant numbers and interaction types are asserted without any quantitative metrics, baseline comparisons, error bars, or ablation results; this absence makes it impossible to assess whether the pipeline actually preserves collaborative dynamics or merely removes local artifacts.
  2. [Method (diffusion-based noise optimization framework)] Diffusion-based refinement stage (described in the method): the approach augments single-person motion priors with high-level objectives for contact consistency and semantic alignment, but provides no derivation or validation showing that these objectives encode mutual anticipation and continuous inter-participant adjustment rather than only local constraints; because the priors originate from individual motion data, this gap is load-bearing for the claim that refined motions maintain collaborative dynamics across diverse interaction types.
  3. [Experiments] Experimental results section: the robustness claim across varying numbers of participants and interaction types requires explicit cross-condition quantitative evaluation (e.g., metrics stratified by participant count or interaction category) with statistical significance; without such breakdowns or comparisons to single-person-only baselines, the added interaction objectives' contribution cannot be isolated.
minor comments (2)
  1. [Stage 1] The description of how optimized grasps are 'extended into complete hand-object interaction sequences' lacks implementation details on temporal consistency enforcement.
  2. [Method] Notation for the interaction objectives (e.g., weights on human-object vs. human-human terms) should be formalized with explicit equations to allow reproducibility.

Simulated Author's Rebuttal

3 responses · 0 unresolved

We thank the referee for their constructive feedback. We will revise the manuscript to strengthen the presentation of quantitative results and provide additional clarifications in the method and experiments sections.

read point-by-point responses

  1. Referee: [Abstract] Abstract: the central claims of effectiveness and robustness across participant numbers and interaction types are asserted without any quantitative metrics, baseline comparisons, error bars, or ablation results; this absence makes it impossible to assess whether the pipeline actually preserves collaborative dynamics or merely removes local artifacts.

    Authors: The abstract summarizes the paper's claims, while detailed quantitative support, including metrics, baselines, and ablations, is provided in the experiments section. We agree this could be better highlighted and will update the abstract to include key quantitative findings on effectiveness and robustness. revision: yes

  2. Referee: [Method (diffusion-based noise optimization framework)] Diffusion-based refinement stage (described in the method): the approach augments single-person motion priors with high-level objectives for contact consistency and semantic alignment, but provides no derivation or validation showing that these objectives encode mutual anticipation and continuous inter-participant adjustment rather than only local constraints; because the priors originate from individual motion data, this gap is load-bearing for the claim that refined motions maintain collaborative dynamics across diverse interaction types.

    Authors: The interaction objectives are specifically designed to couple the motions of multiple participants through shared contact and semantic terms, thereby encoding collaborative dynamics beyond local constraints. We will add further explanation and examples in the revised method section to validate how these objectives promote inter-participant adjustment. revision: partial

  3. Referee: [Experiments] Experimental results section: the robustness claim across varying numbers of participants and interaction types requires explicit cross-condition quantitative evaluation (e.g., metrics stratified by participant count or interaction category) with statistical significance; without such breakdowns or comparisons to single-person-only baselines, the added interaction objectives' contribution cannot be isolated.

    Authors: We have evaluated on data with varying participant counts and interaction types, with overall results supporting robustness. To address the request for stratified evaluation, we will include additional breakdowns by participant number and interaction type in the experiments section, along with comparisons to single-person baselines and statistical analysis. revision: yes

Circularity Check

0 steps flagged

No circularity in derivation; method uses external priors plus new objectives

full rationale

The paper describes a two-stage pipeline: (1) optimization to produce physically plausible grasps from noisy body input, (2) diffusion-based refinement of full-body motion that starts from single-person motion priors and augments them with human-object and human-human interaction objectives. No equation or step equates a claimed output (enhanced collaborative motion) to its inputs by construction, nor does any load-bearing claim rest on a self-citation chain or fitted parameter renamed as prediction. The priors are external; the added objectives are presented as novel constraints. The experimental claims rest on qualitative/quantitative evaluation against capture artifacts rather than on any self-referential reduction. This is the normal non-circular case for a methods paper.

Axiom & Free-Parameter Ledger

1 free parameters · 1 axioms · 0 invented entities

Review performed on abstract only; ledger entries are inferred from the high-level description of the method.

free parameters (1)
  • weights on interaction objectives
    The diffusion optimization introduces objectives to encode human-object and human-human information; their relative strengths are not derived from first principles and must be set or tuned.
axioms (1)
  • domain assumption Single-person motion priors remain a useful base when augmented with interaction terms for collaborative multi-person scenarios
    Invoked when the second stage adapts single-person diffusion models to MHOI data.

pith-pipeline@v0.9.1-grok · 5807 in / 1383 out tokens · 40431 ms · 2026-06-27T01:19:07.872081+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

108 extracted references · 19 canonical work pages · 4 internal anchors

  1. [1]

    ACM Trans

    Object Motion Guided Human Motion Synthesis , author=. ACM Trans. Graph. , volume=

  2. [2]

    ECCV , year=

    Controllable human-object interaction synthesis , author=. ECCV , year=

  3. [3]

    ICCV , year=

    Human-object interaction from human-level instructions , author=. ICCV , year=

  4. [4]

    arXiv preprint arXiv:2506.15625 , year=

    Hoidini: Human-object interaction through diffusion noise optimization , author=. arXiv preprint arXiv:2506.15625 , year=

    work page arXiv

  5. [5]

    Xu, Sirui and Li, Zhengyuan and Wang, Yu-Xiong and Gui, Liang-Yan , booktitle=

  6. [6]

    arXiv preprint arXiv:2403.11237 , year=

    FORCE: Physics-aware Human-object Interaction , author=. arXiv preprint arXiv:2403.11237 , year=

    work page arXiv

  7. [7]

    ECCV , year =

    COUCH: Towards Controllable Human-Chair Interactions , author =. ECCV , year =

  8. [8]

    CVPR , year=

    Sapien: A simulated part-based interactive environment , author=. CVPR , year=

  9. [9]

    , author=

    Neural state machine for character-scene interactions. , author=. ACM Trans. Graph. , volume=

  10. [10]

    ICCV , year =

    Stochastic Scene-Aware Motion Prediction , author =. ICCV , year =

  11. [11]

    AAAI , year=

    Learning to sit: Synthesizing human-chair interactions via hierarchical control , author=. AAAI , year=

  12. [12]

    ACM Trans

    Model predictive control with a visuomotor system for physics-based character animation , author=. ACM Trans. Graph. , volume=

  13. [13]

    ACM Trans

    Learning to use chopsticks in diverse gripping styles , author=. ACM Trans. Graph. , volume=

  14. [14]

    ACM Trans

    Catch & Carry: reusable neural controllers for vision-guided whole-body tasks , author=. ACM Trans. Graph. , volume=

  15. [15]

    2022 , booktitle =

    Lee, Seunghwan and Chang, Phil Sik and Lee, Jehee , title =. 2022 , booktitle =

    2022

  16. [16]

    ACM SIGGRAPH 2023 Conference Proceedings , year=

    Synthesizing physical character-scene interactions , author=. ACM SIGGRAPH 2023 Conference Proceedings , year=

    2023

  17. [17]

    ACM SIGGRAPH 2023 Conference Proceedings , year=

    Pmp: Learning to physically interact with environments using part-wise motion priors , author=. ACM SIGGRAPH 2023 Conference Proceedings , year=

    2023

  18. [18]

    CVPR , year=

    Circle: Capture in rich contextual environments , author=. CVPR , year=

  19. [19]

    ECCV , year=

    Gimo: Gaze-informed human motion prediction in context , author=. ECCV , year=

  20. [20]

    CVPR , year=

    Scaling up dynamic human-scene interaction modeling , author=. CVPR , year=

  21. [21]

    arXiv preprint arXiv:2406.19353 , year=

    Core4d: A 4d human-object-human interaction dataset for collaborative object rearrangement , author=. arXiv preprint arXiv:2406.19353 , year=

    work page arXiv

  22. [22]

    CVPR , year=

    Behave: Dataset and method for tracking human object interactions , author=. CVPR , year=

  23. [23]

    arXiv preprint arXiv:2401.10232 , year=

    Parahome: Parameterizing everyday home activities towards 3d generative modeling of human-object interactions , author=. arXiv preprint arXiv:2401.10232 , year=

    work page arXiv

  24. [24]

    ECCV , year=

    Nymeria: A massive collection of multimodal egocentric daily motion in the wild , author=. ECCV , year=

  25. [25]

    arXiv preprint arXiv:2404.00299 , year=

    HOI-M3: Capture Multiple Humans and Objects Interaction within Contextual Environment , author=. arXiv preprint arXiv:2404.00299 , year=

    work page arXiv

  26. [26]

    Proceedings of the ACM SIGGRAPH/Eurographics Symposium on Computer Animation , year=

    Tiling motion patches , author=. Proceedings of the ACM SIGGRAPH/Eurographics Symposium on Computer Animation , year=

  27. [27]

    ACM SIGGRAPH 2006 Papers , year=

    Motion patches: building blocks for virtual environments annotated with motion data , author=. ACM SIGGRAPH 2006 Papers , year=

    2006

  28. [28]

    ACM Transactions on Graphics (TOG) , year=

    Interaction patches for multi-character animation , author=. ACM Transactions on Graphics (TOG) , year=

  29. [29]

    Proceedings of the 2007 ACM symposium on Virtual reality software and technology , year=

    Simulating competitive interactions using singly captured motions , author=. Proceedings of the 2007 ACM symposium on Virtual reality software and technology , year=

    2007

  30. [30]

    Proceedings of the 2008 Symposium on interactive 3D Graphics and Games , year=

    Simulating interactions of avatars in high dimensional state space , author=. Proceedings of the 2008 Symposium on interactive 3D Graphics and Games , year=

    2008

  31. [31]

    IEEE TVCG , year=

    Simulating multiple character interactions with collaborative and adversarial goals , author=. IEEE TVCG , year=

  32. [32]

    Proceedings of the 2006 ACM SIGGRAPH/Eurographics symposium on Computer animation , year=

    Composition of complex optimal multi-character motions , author=. Proceedings of the 2006 ACM SIGGRAPH/Eurographics symposium on Computer animation , year=

    2006

  33. [33]

    European Conference on Computer Vision , year=

    Remos: 3d motion-conditioned reaction synthesis for two-person interactions , author=. European Conference on Computer Vision , year=

  34. [34]

    IJCV , year=

    Intergen: Diffusion-based multi-human motion generation under complex interactions , author=. IJCV , year=

  35. [35]

    CVPR , year=

    Inter-x: Towards versatile human-human interaction analysis , author=. CVPR , year=

  36. [36]

    ACM Transactions on Graphics (TOG) , year=

    Neural animation layering for synthesizing martial arts movements , author=. ACM Transactions on Graphics (TOG) , year=

  37. [37]

    arXiv preprint arXiv:2303.01418 , year=

    Human motion diffusion as a generative prior , author=. arXiv preprint arXiv:2303.01418 , year=

    work page arXiv

  38. [38]

    CVPR , year=

    Towards social artificial intelligence: Nonverbal social signal prediction in a triadic interaction , author=. CVPR , year=

  39. [39]

    ACM Transactions on Graphics (TOG) , year=

    Generating and ranking diverse multi-character interactions , author=. ACM Transactions on Graphics (TOG) , year=

  40. [40]

    ACM Transactions on Graphics (TOG) , year=

    Control strategies for physically simulated characters performing two-player competitive sports , author=. ACM Transactions on Graphics (TOG) , year=

  41. [41]

    ICCV , year =

    Locomotion-Action-Manipulation: Synthesizing Human-Scene Interactions in Complex 3D Environments , author =. ICCV , year =

  42. [42]

    CVPR , year=

    Synthesizing long-term 3d human motion and interaction in 3d scenes , author=. CVPR , year=

  43. [43]

    CVPR , year=

    Towards Diverse and Natural Scene-aware 3D Human Motion Synthesis , author=. CVPR , year=

  44. [44]

    ECCV , year=

    Long-term human motion prediction with scene context , author=. ECCV , year=

  45. [45]

    CVPR , year=

    Scene-aware Generative Network for Human Motion Synthesis , author=. CVPR , year=

  46. [46]

    NeurIPS , year=

    Humanise: Language-conditioned human motion generation in 3d scenes , author=. NeurIPS , year=

  47. [47]

    ACM SIGGRAPH Asia 2024 Conference Proceedings , year=

    Autonomous character-scene interaction synthesis from text instruction , author=. ACM SIGGRAPH Asia 2024 Conference Proceedings , year=

    2024

  48. [48]

    arXiv preprint arXiv:2503.19901 , year=

    TokenHSI: Unified Synthesis of Physical Human-Scene Interactions through Task Tokenization , author=. arXiv preprint arXiv:2503.19901 , year=

    work page arXiv

  49. [49]

    ACM SIGGRAPH 2023 Conference Proceedings , year=

    Simulation and retargeting of complex multi-character interactions , author=. ACM SIGGRAPH 2023 Conference Proceedings , year=

    2023

  50. [50]

    ACM SIGGRAPH 2010 papers , year=

    Spatial relationship preserving character motion adaptation , author=. ACM SIGGRAPH 2010 papers , year=

    2010

  51. [51]

    Denoising Diffusion Implicit Models

    Denoising diffusion implicit models , author=. arXiv preprint arXiv:2010.02502 , year=

    work page internal anchor Pith review Pith/arXiv arXiv 2010

  52. [52]

    Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

    A cross-dataset study for text-based 3D human motion retrieval , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

  53. [53]

    Human Motion Diffusion Model

    Human motion diffusion model , author=. arXiv preprint arXiv:2209.14916 , year=

    work page internal anchor Pith review Pith/arXiv arXiv

  54. [54]

    CVPR , year=

    Optimizing diffusion noise can serve as universal motion priors , author=. CVPR , year=

  55. [55]

    CVPR , year=

    Avatars grow legs: Generating smooth human motion from sparse tracking inputs with diffusion model , author=. CVPR , year=

  56. [56]

    ACM Transactions on Graphics (TOG) , year=

    Robust solving of optical motion capture data by denoising , author=. ACM Transactions on Graphics (TOG) , year=

  57. [57]

    arXiv preprint arXiv:2505.01425 , year=

    GENMO: A GENeralist Model for Human MOtion , author=. arXiv preprint arXiv:2505.01425 , year=

    work page arXiv

  58. [58]

    ICCV , year=

    Learning motion priors for 4d human body capture in 3d scenes , author=. ICCV , year=

  59. [59]

    ICCV , year=

    Humor: 3d human motion model for robust pose estimation , author=. ICCV , year=

  60. [60]

    ICCV , year =

    Shi, Mingyi and Starke, Sebastian and Ye, Yuting and Komura, Taku and Won, Jungdam , title =. ICCV , year =

  61. [61]

    CVPR , year =

    Zhang, Siwei and Bhatnagar, Bharat Lal and Xu, Yuanlu and Winkler, Alexander and Kadlecek, Petr and Tang, Siyu and Bogo, Federica , title =. CVPR , year =

  62. [62]

    CVPR , year=

    Decoupling Human and Camera Motion from Videos in the Wild , author=. CVPR , year=

  63. [63]

    2025 , journal=

    GENMO: A GENeralist Model for Human MOtion , author=. 2025 , journal=

    2025

  64. [64]

    ACM Transactions on Graphics (TOG) , year=

    Phase-functioned neural networks for character control , author=. ACM Transactions on Graphics (TOG) , year=

  65. [65]

    2022 , journal =

    Starke, Sebastian and Mason, Ian and Komura, Taku , title =. 2022 , journal =

    2022

  66. [66]

    NeurIPS , year=

    Nemf: Neural motion fields for kinematic animation , author=. NeurIPS , year=

  67. [67]

    CVPR , year=

    Ego-Body Pose Estimation via Ego-Head Pose Estimation , author=. CVPR , year=

  68. [68]

    ACM Transactions on Graphics (TOG) , year=

    Physics-based character controllers using conditional vaes , author=. ACM Transactions on Graphics (TOG) , year=

  69. [69]

    ACM Transactions on Graphics (TOG) , year=

    Character controllers using motion vaes , author=. ACM Transactions on Graphics (TOG) , year=

  70. [70]

    ICCV , year=

    Guided motion diffusion for controllable human motion synthesis , author=. ICCV , year=

  71. [71]

    ICLR , year=

    DartControl: A diffusion-based autoregressive motion model for real-time text-driven motion control , author=. ICLR , year=

  72. [72]

    arXiv preprint arXiv:2405.11126 , year=

    Flexible Motion In-betweening with Diffusion Models , author=. arXiv preprint arXiv:2405.11126 , year=

    work page arXiv

  73. [73]

    ICLR , year=

    OmniControl: Control Any Joint at Any Time for Human Motion Generation , author=. ICLR , year=

  74. [74]

    CVPR , year =

    Mocap Everyone Everywhere: Lightweight Motion Capture With Smartwatches and a Head-Mounted Camera , author =. CVPR , year =

  75. [75]

    Pavlakos, Georgios and Choutas, Vasileios and Ghorbani, Nima and Bolkart, Timo and Osman, Ahmed A. A. and Tzionas, Dimitrios and Black, Michael J. , booktitle =. Expressive Body Capture:

  76. [76]

    ShapeNet: An Information-Rich 3D Model Repository

    Shapenet: An information-rich 3d model repository , author=. arXiv preprint arXiv:1512.03012 , year=

    work page internal anchor Pith review Pith/arXiv arXiv

  77. [77]

    NeurIPS , year=

    Denoising diffusion probabilistic models , author=. NeurIPS , year=

  78. [78]

    CVPR , year=

    Contactopt: Optimizing contact to improve grasps , author=. CVPR , year=

  79. [79]

    Dexgraspnet: A large-scale robotic dexterous grasp dataset for general objects based on simulation

    Dexgraspnet: A large-scale robotic dexterous grasp dataset for general objects based on simulation , author=. arXiv preprint arXiv:2210.02697 , year=

    work page arXiv

  80. [80]

    IEEE Robotics and Automation Letters , year=

    Synthesizing diverse and physically stable grasps with arbitrary hand structures using differentiable force closure estimator , author=. IEEE Robotics and Automation Letters , year=

Showing first 80 references.