Pith. sign in

REVIEW 3 major objections 5 minor 3 cited by

SMGDiff: Soccer Motion Generation using diffusion probabilistic models

T0 review · 3 major / 5 minor · reviewed 2026-08-12 · deepseek-v4-flash

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

Pith's one-line read SMGDiff, a two-stage diffusion framework, synthesizes real-time, user-controllable soccer motions and reports better motion quality and condition alignment than the compared controllers.

desk verdict The dataset and pipeline are solid contributions, but the SOTA claim rests on a baseline set that omits CAMDM, the method the paper directly builds on. read the letter →

arxiv 2411.16216 v1 pith:WELYAQQR submitted 2024-11-25 cs.CV

Hongdi Yang , Chengyang Li , Zhenxuan Wu , Gaozheng Li , Jingya Wang , Jingyi Yu , Zhuo Su , Lan Xu This is my paper · ORCID
classification cs.CV
keywords soccermotiongenerationdiffusionprobabilisticmodelsreal-timecharactercontrolhuman-objectinteractionball-footcontactcapturedatasetautoregressivetrajectoryconditioning
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

SMGDiff's central claim is that realistic soccer animation can be generated in real time from coarse user controls by splitting the task into two stages: a lightweight trajectory generator that turns keyboard-style inputs into a smooth future path, and a transformer-based autoregressive diffusion model that produces the body, ball, and contact states along that path. The paper adds a contact-guidance pass at inference to sharpen ball-foot interactions, and contributes Soccer-X, a mocap dataset of over 1.08 million frames covering six soccer skill categories from 30 players. On its own test set the system reports the lowest FID and foot-sliding distance and the highest skill accuracy among the compared real-time controllers. A sympathetic reader would take the paper as establishing that diffusion-based generation, rather than example-based matching or physics-based control, is a viable route to interactive soccer animation.

What carries the argument

The carrying object is the two-stage pipeline itself. Stage one is a single-step diffusion transformer: conditioned on skill label, target trajectory point, past trajectory, and injected Gaussian noise, it predicts a future ground-plane trajectory, and a temporal blending strategy prevents abrupt turns. Stage two is a transformer-based autoregressive diffusion model that takes as condition the skill label, the refined future trajectory, and past soccer motion, and predicts the future motion state $x_i = \{h, b, c\}$ with human pose $h$, ball state $b$, and binary contact labels $c$; it is trained with the denoising diffusion probabilistic model (DDPM) simple objective plus auxiliary position, velocity, and foot-contact losses. The contact-guidance module then detects a ball-foot contact when ball acceleration exceeds $\tau_a = 2\,\mathrm{m/s^2}$ (Eq. 10), picks the lifted foot nearest the ball, and uses a distance-based contact loss with the DSG spherical-Gaussian guidance rule to steer the last two denoising steps. The Soccer-X dataset supplies the learned distribution: 1.08 million frames, six skill classes, 30 players.

What would settle it

Compute, on the held-out split of Soccer-X, precision and recall of the Eq. (10) contact detector ($\|b_a\| > 2\,\mathrm{m/s^2}$) against the recorded binary ball-foot contact labels; if many true contacts have acceleration below threshold or many non-contacts exceed it, the contact-guidance signal is driven by an unvalidated proxy, and the FID improvement attributed to the module needs a different explanation.

Watch

Extended reading notes

Core claim

The discovery the paper argues for is a generation recipe: coarse user commands are converted into a global trajectory by a single-step diffusion transformer, and that trajectory conditions an autoregressive diffusion model that generates future soccer motion jointly for the SMPL body, the ball, and binary foot-ground and foot-ball contacts. The ball-foot contact detail is then improved at inference by a contact-loss guidance that detects contact from a threshold on ball acceleration and pulls the nearest lifted foot toward the ball. The paper reports that this recipe yields FID 0.1813, foot-sliding distance 0.8543 m, and skill accuracy 93.3% on its evaluation split, against best baseline values of 0.2494, 1.0678 m, and 73.3%, and that the full pipeline runs at interactive speed with 8 denoising steps.

Load-bearing premise

The load-bearing premise is that a ball-foot contact is happening whenever the generated ball's acceleration exceeds 2 m/s², and that using that inferred contact as a target for the guidance loss makes motions more realistic; the paper does not validate this detector against the ground-truth contact labels it collects.

Editorial extensions

If this is right

  • On the paper's reported evaluation, SMGDiff reaches FID 0.1813, foot sliding 0.8543 m, and skill accuracy 93.3%, while the best compared baseline reaches 0.2494, 1.0678 m, and 73.3%.
  • The system runs at interactive speeds with 8 denoising steps and contact guidance confined to the last two steps, with a reported inference time of 12 ms for a generated segment.
  • The Soccer-X dataset, with 1.08 million frames across six skill categories and 30 players, provides a new training and evaluation resource for data-driven soccer motion generation.
  • Contact guidance specifically reduces missed ball-foot contacts when the ball changes direction, a failure mode visible in the compared controllers.

Reading between the lines

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

  • An extension the paper leaves open is to replace the acceleration-threshold contact detector with a learned classifier trained on the same binary contact labels the dataset contains, and to compare FID and skill accuracy under both detectors.
  • The trajectory-then-motion decomposition is generic enough that the same recipe could transfer to other fast object-interaction animation tasks such as basketball or tennis, where user control is a coarse target rather than a full pose sequence.
  • Because the contact guidance is enabled only for the final two denoising steps, varying both the number of guided steps and their position in the schedule would isolate whether the reported FID gain comes from contact correctness itself or from the extra gradient signal at low noise.
  • The reported real-time figure covers model inference and network communication; an end-to-end measurement that includes rendering would quantify the latency a player actually perceives.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 5 minor

Summary. SMGDiff is a two-stage framework for real-time, user-controllable soccer motion generation. The first stage converts coarse user controls (skill label, target point, and past trajectory) into smooth global trajectories using a single-step diffusion model with temporal blending borrowed from CAMDM. The second stage trains a transformer-based autoregressive diffusion model to generate future human, ball, and contact states conditioned on trajectory, skill, and past motion; during inference, a contact guidance module applies a custom loss in the final denoising steps to improve ball-foot contact. The authors contribute Soccer-X, a 1.08M-frame motion capture dataset with six soccer skill categories, and report experiments on their own test set against LMP, MANN-DP, and CM, achieving FID 0.1813 and Skill Accuracy 93.3%, with ablations and runtime analysis showing that 8 denoise steps and contact guidance on the last two steps are preferred.

Significance. If validated, the paper would make a useful applied contribution: a large soccer motion dataset, a real-time interactive pipeline, and a concrete way to inject ball-foot contact into autoregressive diffusion. The dataset size and the live Unity demo are tangible strengths, and the use of standard diffusion losses and a single-step trajectory generator is technically sound. The core concern is that the central SOTA claim is not yet supported: the closest diffusion-based real-time controller, CAMDM, is absent from the comparison, and the contact-guidance module's target is inferred from the very motion it is supposed to correct. These are fixable with additional experiments and validation, but they are load-bearing.

major comments (3)
  1. [Sec. 5.1, Table 1] The central claim that SMGDiff 'significantly outperforms existing methods' is not established because CAMDM [5], the method on which SMGDiff directly builds, is never included as a baseline. SMGDiff's second stage is described as 'a transformer-based autoregressive diffusion model [5]' (Sec. 3.2), its trajectory blending uses CAMDM's HFTE strategy (Sec. 3.1), and its skill-accuracy evaluation follows [5] (Sec. 5). Yet Table 1 compares only LMP, MANN-DP, and CM, all non-diffusion controllers predating CAMDM. Since CAMDM is the relevant state of the art for real-time controllable diffusion-based character motion, the reported FID 0.1813 and Skill Acc. 93.3% do not demonstrate superiority over the architecture on which SMGDiff is built. A CAMDM baseline trained on Soccer-X with the same representation and condition inputs must be added, or the 'outperforms existing methods' claim must be scaled back.
  2. [Sec. 3.3, Eqs. (10)-(12)] The contact guidance target is self-referential and unvalidated. Eq. (10) infers the contact indicator ĉb from the generated ball acceleration exceeding a hand-set threshold τa=2 m/s², and Eq. (12) then uses that inferred contact to guide the same generated motion. Because the guidance signal is derived from the variable being optimized, it cannot provide independent evidence of contact; if the acceleration heuristic is inaccurate, guidance will actively distort motion around non-contact frames. The dataset already contains ground-truth contact labels c (defined in Sec. 3), so the heuristic can be validated or replaced by a learned contact predictor. The CGM ablation in Sec. 5.2 is therefore inconclusive as evidence of improved physical contact.
  3. [Sec. 5, Tables 1-4] The empirical support for 'significantly outperforms' is statistically weak. No error bars, confidence intervals, or significance tests are reported; the numbers appear to be single-run point estimates. In addition, evaluation is confined to the authors' own Soccer-X test set, and for the Shoot category the ball trajectories were reconstructed in Unity rather than captured (Sec. 4), which may bias contact-related metrics in that category. The paper should report per-category metrics with variance across seeds, perform a significance test, and quantify the effect of the Unity-based Shoot reconstruction on contact accuracy.
minor comments (5)
  1. [Sec. 3.2] The text attributes the simple objective to 'Go et al. [15]', but reference [15] is Ho et al.; please correct the citation.
  2. [Tables 1 and 2] The Diversity and FID values for 'Ours' differ substantially between Table 1 (Div. 0.6177, FID 0.1813) and Table 2 (Div. 2.6925, FID 0.3580) under the same model. The paper should explain why the evaluation protocol differs; as written, the ablation gains are hard to interpret.
  3. [Sec. 4] The contribution bullet says 'Our data and model will be disseminated to the community' and Sec. 4 calls Soccer-X a 'large-scale open-source soccer dataset', but no dataset release URL or licensing information is provided; please include these details.
  4. [References] References [15] and [16] are the same paper (Ho et al., DDPM) with different numbering cues; please deduplicate.
  5. [Fig. 2] The pipeline figure uses 'SMDM' without defining the acronym in the text or caption; please define it at first use.

Circularity Check

1 steps flagged · score 4.0 of 10

Contact guidance defines its own contact targets from the generated ball acceleration, making the claimed contact refinement partly self-referential; the overall SOTA claim still rests on independently computed FID and skill metrics.

  1. self definitional [Sec. 3.3, Eq. (10) and Eq. (12); Fig. 6 caption]
    "When the ball's acceleration exceeds a certain threshold, we infer that the ball is influenced by both ground friction and the force exerted by the foot. In this case, we confirm that contact exists between the ball and foot joints, denoted as ĉb. ĉb = I(||ba|| > τa), (10) ... Finally, our contact loss function for guidance is L = sum_i d_i · I(d_i > τd) · ĉi_b / (I(d_i > τd) + δ), (12) ... Contact frames represent points where the ball’s trajectory shifts."

    The guidance target is generated internally: Eq. (10) defines the contact label ĉb from the model's own predicted ball acceleration, and Eq. (12) uses this same ĉb to select the frames where the contact loss is applied and to weight its magnitude. The generated ball trajectory is thus both the source of the 'contact ground truth' and the object being optimized toward it. No external ground-truth contact label (the dataset already contains c_b) is used to validate the threshold, and the qualitative evaluation in Fig. 6 defines contact frames as 'points where the ball's trajectory shifts,' i.e., by the same heuristic. The claimed contact refinement is therefore a self-consistency regularizer rather than an independently measured improvement in contact accuracy.

full rationale

The central derivation chain is otherwise self-contained: the trajectory generation model, the autoregressive diffusion backbone, and the FID/foot-sliding/skill-accuracy comparisons are computed against conventional baselines using the same test distribution, so the main 'significantly outperforms' claim does not reduce to its own inputs. The contact guidance module is the one place where the paper's own equation defines the quantity to be optimized: Eq. (10) infers contact from the generated ball acceleration, and Eq. (12) optimizes the same sample toward that inferred contact. This is a genuine self-referential step, but it affects a single module rather than the whole framework. The omission of CAMDM as a baseline, while a real evaluation concern, is not circularity: CAMDM is cited as the source of the autoregressive diffusion and HFTE blending, but the soccer-specific contribution is not reduced to that citation. Self-citations such as the DSG guidance update in Eqs. (13)-(15) are also not load-bearing in a circular sense, since the guidance framework is used as a standard tool rather than as the justification for the contact result. Overall, the paper has one partially self-definitional contact-evaluation loop, but the core SOTA claim retains independent content, so a moderate score is appropriate.

Assumptions & free parameters 6 free parameters · 4 assumptions · 0 invented entities

The central contribution rests on a small number of hand-set thresholds and heuristics (τa, τd, wd, wr) and on two domain assumptions about contact physics, plus Unity-simulated ball paths for shooting. No new physical entities are introduced; the claimed novelty is in the dataset and the guidance heuristic.

free parameters (6)
  • τa (contact acceleration threshold) = 2 m/s²
    Hand-set in Eq. (10); determines when the model infers a ball-foot contact from predicted ball acceleration.
  • τd (contact distance threshold) = 0.1 m
    Hand-set in Eq. (12); gates the contact loss so it only pushes feet within 0.1 m.
  • wd (lifted foot penalty) = 2
    Hand-set in Eq. (11); biases contact joint selection toward the lifted foot.
  • wr (guidance rate) = 0.5
    Hand-set in Eq. (14); balances unconditional sampling direction and guidance direction.
  • r (ball control radius) = 2 m
    Defines the ball control weight in Eq. (1) and converts global ball position to relative; a representation parameter.
  • Loss weights λpos, λvel, λfoot
    Weights for auxiliary losses in Eq. (9) are not reported in the main text, so the training objective is incompletely specified.
assumptions (4)
  • ad hoc to paper Ball-foot contact can be inferred from ball acceleration exceeding 2 m/s² (Eq. 10).
    The paper justifies it with 'laws of mechanics' but the threshold is hand-picked and the rule is not validated against captured contact labels.
  • domain assumption The lifted foot is the most plausible contact joint for ball-foot interaction (Eq. 11).
    A reasonable soccer heuristic, but it may fail for toe taps or sole rolls where the planted foot touches the ball.
  • domain assumption Ground friction gives the ball a small, constant acceleration, so deviations from that baseline indicate foot contact.
    Used to justify Eq. (10); ignores spin, bounces, and non-constant friction, and is not empirically calibrated.
  • domain assumption Shoot ball trajectories reconstructed in Unity physics are realistic enough to train and evaluate on.
    The dataset section states full ball trajectories were rebuilt in Unity due to indoor space limits; this introduces simulated data into the training distribution without a validation of its fidelity.

how reviews work

0 comments
Cite this review

Pith. "Pith review of SMGDiff: Soccer Motion Generation using diffusion probabilistic models." pith.science (2026). https://pith.science/paper/WELYAQQR

@misc{pith2026241116216,
  author       = {Pith},
  title        = {Pith review of: SMGDiff: Soccer Motion Generation using diffusion probabilistic models},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/WELYAQQR}},
  note         = {Machine review of arXiv:2411.16216}
}
read the original abstract

Soccer is a globally renowned sport with significant applications in video games and VR/AR. However, generating realistic soccer motions remains challenging due to the intricate interactions between the human player and the ball. In this paper, we introduce SMGDiff, a novel two-stage framework for generating real-time and user-controllable soccer motions. Our key idea is to integrate real-time character control with a powerful diffusion-based generative model, ensuring high-quality and diverse output motion. In the first stage, we instantly transform coarse user controls into diverse global trajectories of the character. In the second stage, we employ a transformer-based autoregressive diffusion model to generate soccer motions based on trajectory conditioning. We further incorporate a contact guidance module during inference to optimize the contact details for realistic ball-foot interactions. Moreover, we contribute a large-scale soccer motion dataset consisting of over 1.08 million frames of diverse soccer motions. Extensive experiments demonstrate that our SMGDiff significantly outperforms existing methods in terms of motion quality and condition alignment.

Figures

Figures reproduced from arXiv: 2411.16216 by the authors.

Figure 1
Figure 1. Our method, SMGDiff, enables users to control soccer motions based on character displacement and soccer skill, simulating an [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Pipeline of SMGDiff. Our framework consists of two stages: In the trajectory generation stage, we transform soccer skill label [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. The top section exhibits selected highlights of our dataset. [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figures from the paper (2 more)
Figure 4
Figure 4. Figure 4: Qualitative comparison between our method and baseline methods including LMP [ [PITH_FULL_IMAGE:figures/full_fig_p006_4.png]
Figure 6
Figure 6. Figure 6: Qualitative evaluation of the Contact Guidance Module [PITH_FULL_IMAGE:figures/full_fig_p007_6.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 3 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. FixTalk: Taming Identity Leakage for High-Quality Talking Head Generation in Extreme Cases

    cs.CV 2025-07 conditional novelty 5.0 of 10

    FixTalk adds two modules to a real-time GAN talking-head model, decoupling identity from motion to stop identity leakage while using a memory to recover details and reduce artifacts.

  2. EnvPoser: Environment-aware Realistic Human Motion Estimation from Sparse Observations with Uncertainty Modeling

    cs.CV 2024-12 conditional novelty 5.0 of 10

    Using a pre-scanned environment point cloud as context reduces ambiguity in full-body pose estimation from sparse VR trackers, improving accuracy over previous methods.

  3. Multimodal Generative AI with Autoregressive LLMs for Human Motion Understanding and Generation: A Way Forward

    cs.CV 2025-05 conditional novelty 3.0 of 10

    A survey paper reviews multimodal generative AI and autoregressive LLMs for text-driven human motion generation, with comparative tables of models, datasets, and metrics.

Reference graph

Works this paper leans on

81 extracted references · 72 canonical work pages · cited by 3 Pith papers

  1. [5]

    Taming diffusion probabilistic models for character control

    Rui Chen, Mingyi Shi, Shaoli Huang, Ping Tan, Taku Komura, and Xuelin Chen. Taming diffusion probabilistic models for character control. In SIGGRAPH, 2024. 1, 2, 4, 6, 7, 8

    work page 2024

  2. [1]

    Listen, denoise, action! audio-driven motion synthesis with diffusion models

    Simon Alexanderson, Rajmund Nagy, Jonas Beskow, and Gustav Eje Henter. Listen, denoise, action! audio-driven motion synthesis with diffusion models. ACM Trans. Graph.,

  3. [2]

    Gesturediffuclip: Gesture diffusion model with clip latents

    Tenglong Ao, Zeyi Zhang, and Libin Liu. Gesturediffuclip: Gesture diffusion model with clip latents. ACM Transactions on Graphics (TOG), 42(4):1–18, 2023. 2

    work page 2023

  4. [3]

    Okan Arikan and David A. Forsyth. Interactive motion gener- ation from examples. ACM Transactions on Graphics (TOG),

    work page

  5. [4]

    Diffusiontalker: Person- alization and acceleration for speech-driven 3d face diffuser

    Peng Chen, Xiaobao Wei, Ming Lu, Yitong Zhu, Naiming Yao, Xingyu Xiao, and Hui Chen. Diffusiontalker: Person- alization and acceleration for speech-driven 3d face diffuser. arXiv preprint arXiv:2311.16565, 2023. 2

    arXiv 2023

  6. [6]

    Executing your commands via motion diffusion in latent space

    Xin Chen, Biao Jiang, Wen Liu, Zilong Huang, Bin Fu, Tao Chen, and Gang Yu. Executing your commands via motion diffusion in latent space. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages 18000–18010, 2023. 2

    2023

  7. [7]

    Motion recommendation for online char- acter control

    Kyungmin Cho, Chaelin Kim, Jungjin Park, Joonkyu Park, and Junyong Noh. Motion recommendation for online char- acter control. ACM Transactions on Graphics (TOG), 2021. 2

    work page 2021

  8. [8]

    Motion matching and the road to next-gen ani- mation

    Simon Clavet. Motion matching and the road to next-gen ani- mation. In Proceedings of the Game Developers Conference (GDC), 2016. 2

    work page 2016

Show all 81 references
  1. [9]

    Cg-hoi: Contact-guided 3d human-object interaction generation

    Christian Diller and Angela Dai. Cg-hoi: Contact-guided 3d human-object interaction generation. In 2024 IEEE/CVF Con- ference on Computer Vision and Pattern Recognition (CVPR),

    2024

  2. [10]

    EA SPORTS FC - Official Website

    Electronic Arts. EA SPORTS FC - Official Website. https: //www.ea.com/games/ea-sports-fc, 2024. 1, 2

    2024

  3. [11]

    Recurrent network models for human dynamics

    Katerina Fragkiadaki, Sergey Levine, Panna Felsen, and Jiten- dra Malik. Recurrent network models for human dynamics. In Proceedings of the IEEE International Conference on Com- puter Vision (ICCV), pages 4346–4354. IEEE, 2015. 2

    2015

  4. [12]

    Imos: Intent-driven full- body motion synthesis for human-object interactions.Comput

    Anindita Ghosh, Rishabh Dabral, Vladislav Golyanik, Chris- tian Theobalt, and Philipp Slusallek. Imos: Intent-driven full- body motion synthesis for human-object interactions.Comput. Graph. Forum, 42(2):1–12, 2023. 2

    2023

  5. [13]

    Generating diverse and natural 3d human motions from text

    Chuan Guo, Shihao Zou, Xinxin Zuo, Sen Wang, Wei Ji, Xingyu Li, and Li Cheng. Generating diverse and natural 3d human motions from text. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages 5152–5161, 2022. 1

    2022

  6. [14]

    Synthesizing physical character-scene interactions, 2023

    Mohamed Hassan, Yunrong Guo, Tingwu Wang, Michael Black, Sanja Fidler, and Xue Bin Peng. Synthesizing physical character-scene interactions, 2023. 2

    2023

  7. [15]

    Denoising diffu- sion probabilistic models

    Jonathan Ho, Ajay Jain, and Pieter Abbeel. Denoising diffu- sion probabilistic models. In Advances in Neural Information Processing Systems, pages 6840–6851. Curran Associates, Inc., 2020. 4

    2020

  8. [16]

    Denoising diffu- sion probabilistic models

    Jonathan Ho, Ajay Jain, and Pieter Abbeel. Denoising diffu- sion probabilistic models. In Advances in Neural Information Processing Systems, 2020. 2

    2020

  9. [17]

    Denoising diffu- sion probabilistic models

    Jonathan Ho, Ajay Jain, and Pieter Abbeel. Denoising diffu- sion probabilistic models. In Advances in Neural Information Processing Systems, pages 6840–6851, 2020. 2

    2020

  10. [18]

    Learning motion manifolds with convolutional autoencoders

    Daniel Holden, Jun Saito, Taku Komura, and Thomas Joyce. Learning motion manifolds with convolutional autoencoders. In SIGGRAPH Asia 2015 Technical Briefs, page 18. ACM,

    2015

  11. [19]

    A deep learning framework for character motion synthesis and editing

    Daniel Holden, Jun Saito, and Taku Komura. A deep learning framework for character motion synthesis and editing. ACM Transactions on Graphics (ToG), 2016. 2

    2016

  12. [20]

    Phase- functioned neural networks for character control

    Daniel Holden, Taku Komura, and Jun Saito. Phase- functioned neural networks for character control. ACM Trans. Graph., 2017. 2

    2017

  13. [21]

    Learned motion matching

    Daniel Holden, Oussama Kanoun, Maksym Perepichka, and Tiberiu Popa. Learned motion matching. ACM Transactions on Graphics (ToG), 2020. 2

    2020

  14. [22]

    Learned motion matching

    Daniel Holden, Oussama Kanoun, Maksym Perepichka, and Tiberiu Popa. Learned motion matching. ACM Trans. Graph.,

  15. [23]

    Shin, and Junyong Noh

    Seokpyo Hong, Daseong Han, Kyungmin Cho, Joseph S. Shin, and Junyong Noh. Physics-based full-body soccer motion control for dribbling and shooting. ACM Trans. Graph., 38 (4), 2019. 1, 2

    2019

  16. [24]

    Diffusion- based generation, optimization, and planning in 3d scenes

    Siyuan Huang, Zan Wang, Puhao Li, Baoxiong Jia, Tengyu Liu, Yixin Zhu, Wei Liang, and Song-Chun Zhu. Diffusion- based generation, optimization, and planning in 3d scenes. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 16750–16761,

  17. [25]

    Guided motion diffusion for con- trollable human motion synthesis

    Korrawe Karunratanakul, Konpat Preechakul, Supasorn Suwa- janakorn, and Siyu Tang. Guided motion diffusion for con- trollable human motion synthesis. In Proceedings of the IEEE/CVF International Conference on Computer Vision , pages 2151–2162, 2023. 2

    2023

  18. [26]

    Opti- mizing diffusion noise can serve as universal motion priors

    Korrawe Karunratanakul, Konpat Preechakul, Emre Aksan, Thabo Beeler, Supasorn Suwajanakorn, and Siyu Tang. Opti- mizing diffusion noise can serve as universal motion priors. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2024. 2

    2024

  19. [27]

    Flame: Free- form language-based motion synthesis & editing

    Jihoon Kim, Jiseob Kim, and Sungjoon Choi. Flame: Free- form language-based motion synthesis & editing. In Proceed- ings of the AAAI Conference on Artificial Intelligence, pages 8255–8263, 2023. 2

    2023

  20. [28]

    Motion graphs

    Lucas Kovar, Michael Gleicher, and Fr´ed´eric Pighin. Motion graphs. ACM Transactions on Graphics, 21(3):473–482, 2002. 2 9

    2002

  21. [29]

    Manfred Lau and James J. Kuffner. Behavior planning for character animation. In Proceedings of the Symposium on Computer Animation (SCA), 2005. 2

    2005

  22. [30]

    Interactive char- acter animation by learning multi-objective control

    Kyungho Lee, Seyoung Lee, and Jehee Lee. Interactive char- acter animation by learning multi-objective control. ACM Transactions on Graphics (TOG), 2018. 2

    2018

  23. [31]

    Learn- ing time-critical responses for interactive character control

    Kyungho Lee, Sehee Min, Sunmin Lee, and Jehee Lee. Learn- ing time-critical responses for interactive character control. ACM Transactions on Graphics (ToG) , 40(4):Article 147,

  24. [32]

    Object motion guided human motion synthe- sis

    Jiaman Li et al. Object motion guided human motion synthe- sis. ACM Trans. Graph., 2023. 2

    2023

  25. [33]

    Task-oriented human-object interactions generation with im- plicit neural representations

    Quanzhou Li, Jingbo Wang, Chen Change Loy, and Bo Dai. Task-oriented human-object interactions generation with im- plicit neural representations. CoRR, abs/2303.13129, 2023. 2

    2023 arXiv

  26. [34]

    Auto-conditioned recurrent networks for ex- tended complex human motion synthesis

    Zimo Li, Yi Zhou, Shuangjiu Xiao, Chong He, Zeng Huang, and Hao Li. Auto-conditioned recurrent networks for ex- tended complex human motion synthesis. arXiv preprint arXiv:1707.05363, 2017. 2

    2017 arXiv

  27. [35]

    Intergen: Diffusion-based multi-human motion gener- ation under complex interactions

    Han Liang, Wenqian Zhang, Wenxuan Li, Jingyi Yu, and Lan Xu. Intergen: Diffusion-based multi-human motion gener- ation under complex interactions. International Journal of Computer Vision, 2023. 2

    2023

  28. [36]

    Omg: Towards open-vocabulary motion generation via mixture of controllers

    Han Liang, Jiacheng Bao, Ruichi Zhang, Sihan Ren, Yuecheng Xu, Sibei Yang, Xin Chen, Jingyi Yu, and Lan Xu. Omg: Towards open-vocabulary motion generation via mixture of controllers. In Proceedings of the IEEE/CVF Con- ference on Computer Vision and Pattern Recognition (CVPR),...

    2024

  29. [37]

    Learning basketball dribbling skills using trajectory optimization and deep reinforcement learning

    Jessica Hodgins Libin Liu. Learning basketball dribbling skills using trajectory optimization and deep reinforcement learning. ACM Transactions on Graphics, August 2018. 2

    2018

  30. [38]

    Siqi Liu, Guy Lever, Zhe Wang, Josh Merel, S. M. Ali Es- lami, Daniel Hennes, Wojciech M. Czarnecki, Yuval Tassa, Shayegan Omidshafiei, Abbas Abdolmaleki, Noah Y . Siegel, Leonard Hasenclever, Luke Marris, Saran Tunyasuvunakool, H. Francis Song, Markus Wulfmeier, Paul Muller, ...

    2021

  31. [39]

    Matthew Loper, Naureen Mahmood, Javier Romero, Gerard Pons-Moll, and Michael J. Black. Smpl: a skinned multi- person linear model. ACM Trans. Graph., 34(6), 2015. 3

    2015

  32. [40]

    Smpl: A skinned multi- person linear model

    Matthew Loper, Naureen Mahmood, Javier Romero, Gerard Pons-Moll, and Michael J Black. Smpl: A skinned multi- person linear model. ACM Transactions on Graphics (TOG), 34(6):1–16, 2015. 6

    2015

  33. [41]

    Motion matching in ’the last of us part ii’, 2021

    Michal Mach and Maksym Zhuravlov. Motion matching in ’the last of us part ii’, 2021. https://www.gdcvault. com/play/1027118/Motion-Matching-in-The- Last. 2

    2021

  34. [42]

    Isaac gym: High performance gpu-based physics simu- lation for robot learning, 2021

    Viktor Makoviychuk, Lukasz Wawrzyniak, Yunrong Guo, Michelle Lu, Kier Storey, Miles Macklin, David Hoeller, Nikita Rudin, Arthur Allshire, Ankur Handa, and Gavriel State. Isaac gym: High performance gpu-based physics simu- lation for robot learning, 2021. 8

    2021

  35. [43]

    Real-time style modelling of human locomotion via feature-wise trans- formations and local motion phases

    Ian Mason, Sebastian Starke, and Taku Komura. Real-time style modelling of human locomotion via feature-wise trans- formations and local motion phases. Proc. ACM Comput. Graph. Interact. Tech., 2022. 2

    2022

  36. [44]

    Catch & carry: Reusable neural controllers for vision-guided whole-body tasks, 2020

    Josh Merel, Saran Tunyasuvunakool, Arun Ahuja, Yuval Tassa, Leonard Hasenclever, Vu Pham, Tom Erez, Greg Wayne, and Nicolas Heess. Catch & carry: Reusable neural controllers for vision-guided whole-body tasks, 2020. 2

    2020

  37. [45]

    GLIDE: Towards photorealistic image gen- eration and editing with text-guided diffusion models

    Alexander Quinn Nichol, Prafulla Dhariwal, Aditya Ramesh, Pranav Shyam, Pamela Mishkin, Bob McGrew, Ilya Sutskever, and Mark Chen. GLIDE: Towards photorealistic image gen- eration and editing with text-guided diffusion models. In In- ternational Conference on Machine Learning ...

    2022

  38. [46]

    NaturalPoint, Inc

    Optitrack. NaturalPoint, Inc. Motion Capture Systems. https://optitrack.com/, 2022. 2, 5

    2022

  39. [47]

    Mcp: Learning composable hierarchical control with multiplicative compositional policies, 2019

    Xue Bin Peng, Michael Chang, Grace Zhang, Pieter Abbeel, and Sergey Levine. Mcp: Learning composable hierarchical control with multiplicative compositional policies, 2019. 1, 2

    2019

  40. [48]

    Amp: adversarial motion priors for styl- ized physics-based character control

    Xue Bin Peng, Ze Ma, Pieter Abbeel, Sergey Levine, and Angjoo Kanazawa. Amp: adversarial motion priors for styl- ized physics-based character control. ACM Transactions on Graphics, 2021. 2

    2021

  41. [49]

    Amp: Adversarial motion priors for styl- ized physics-based character control

    Xue Bin Peng, Ze Ma, Pieter Abbeel, Sergey Levine, and Angjoo Kanazawa. Amp: Adversarial motion priors for styl- ized physics-based character control. ACM Transactions on Graphics (ToG), 40(4):1–20, 2021. 1

    2021

  42. [50]

    Hierarchical text-conditional image genera- tion with clip latents

    Aditya Ramesh, Prafulla Dhariwal, Alex Nichol, Casey Chu, and Mark Chen. Hierarchical text-conditional image genera- tion with clip latents. ArXiv, 2022. 4

    2022

  43. [51]

    Photorealistic text-to-image diffusion models with deep language understanding

    Chitwan Saharia, William Chan, Saurabh Saxena, Lala Li, Jay Whang, Emily L Denton, Kamyar Ghasemipour, Raphael Gontijo Lopes, Burcu Karagol Ayan, Tim Salimans, and et al. Photorealistic text-to-image diffusion models with deep language understanding. In Advances in Neural Info...

    2022

  44. [52]

    Controllable motion diffusion model

    Yi Shi, Jingbo Wang, Xuekun Jiang, and Bo Dai. Controllable motion diffusion model. CoRR, 2023. 8

    2023

  45. [53]

    Denoising diffusion implicit models, 2022

    Jiaming Song, Chenlin Meng, and Stefano Ermon. Denoising diffusion implicit models, 2022. 8

    2022

  46. [54]

    Neural state machine for character-scene interactions

    Sebastian Starke, He Zhang, Taku Komura, and Jun Saito. Neural state machine for character-scene interactions. ACM Trans. Graph., 38(6), 2019. 2

    2019

  47. [55]

    Neural state machine for character-scene interactions

    Sebastian Starke, He Zhang, Taku Komura, and Jun Saito. Neural state machine for character-scene interactions. ACM Transactions on Graphics (ToG), 38(6):209:1–209:10, 2019. 2

    2019

  48. [56]

    Local motion phases for learning multi-contact character movements

    Sebastian Starke, Yiwei Zhao, Taku Komura, and Kazi Za- man. Local motion phases for learning multi-contact character movements. ACM Trans. Graph., 39(4), 2020. 2, 3

    2020

  49. [57]

    Local motion phases for learning multi-contact character movements

    Sebastian Starke, Yiwei Zhao, Taku Komura, and Kazi Za- man. Local motion phases for learning multi-contact character movements. ACM Transactions on Graphics (SIGGRAPH),

  50. [58]

    Neural animation layering for synthesizing martial arts movements

    Sebastian Starke, Yiwei Zhao, Fabio Zinno, and Taku Ko- mura. Neural animation layering for synthesizing martial arts movements. ACM Transactions on Graphics (TOG), 2021. 2 10

    2021

  51. [59]

    Deepphase: Periodic autoencoders for learning motion phase manifolds

    Sebastian Starke, Ian Mason, and Taku Komura. Deepphase: Periodic autoencoders for learning motion phase manifolds. ACM Transactions on Graphics (ToG), 2022. 6, 7

    2022

  52. [60]

    Categorical codebook matching for embodied character controllers

    Sebastian Starke, Paul Starke, Nicky He, Taku Komura, and Yuting Ye. Categorical codebook matching for embodied character controllers. ACM Trans. Graph., 2024. 2, 6, 7

    2024

  53. [61]

    Guy Tevet, Sigal Raab, Brian Gordon, Yonatan Shafir, Daniel Cohen-Or, and Amit H. Bermano. Human motion diffusion model. In International Conference on Learning Representa- tions (ICLR), 2023. 1, 2, 4, 6

    2023

  54. [62]

    Closd: Closing the loop between simulation and diffusion for multi-task character control

    Guy Tevet, Sigal Raab, Setareh Cohan, Daniele Reda, Zhengyi Luo, Xue Bin Peng, Amit H Bermano, and Michiel van de Panne. Closd: Closing the loop between simulation and diffusion for multi-task character control. arXiv preprint arXiv:2410.03441, 2024. 2

    arXiv

  55. [63]

    Karen Liu

    Jonathan Tseng, Rodrigo Castellon, and C. Karen Liu. Edge: Editable dance generation from music. In 2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2023. 1, 2

    2023

  56. [64]

    Omnicontrol: Control any joint at any time for human motion generation

    Yiming Xie, Varun Jampani, Lei Zhong, Deqing Sun, and Huaizu Jiang. Omnicontrol: Control any joint at any time for human motion generation. In The Twelfth International Conference on Learning Representations, 2023. 2

    2023

  57. [65]

    Learning soccer juggling skills with layer-wise mixture-of-experts

    Zhaoming Xie, Sebastian Starke, Hung Yu Ling, and Michiel van de Panne. Learning soccer juggling skills with layer-wise mixture-of-experts. In ACM SIGGRAPH 2022 Conference Proceedings, 2022. 1, 2

    2022

  58. [66]

    Karen Liu

    Zhaoming Xie, Jonathan Tseng, Sebastian Starke, Michiel van de Panne, and C. Karen Liu. Hierarchical planning and control for box loco-manipulation, 2023. 2

    2023

  59. [67]

    Guidance with spherical gaussian con- straint for conditional diffusion

    Lingxiao Yang, Shutong Ding, Yifan Cai, Jingyi Yu, Jingya Wang, and Ye Shi. Guidance with spherical gaussian con- straint for conditional diffusion. In International Conference on Machine Learning, 2024. 5

    2024

  60. [68]

    Controlvae: Model-based learning of generative controllers for physics-based characters

    Heyuan Yao, Zhenhua Song, Baoquan Chen, and Libin Liu. Controlvae: Model-based learning of generative controllers for physics-based characters. ACM Trans. Graph., 2022. 2

    2022

  61. [69]

    Moconvq: Unified physics- based motion control via scalable discrete representations

    Heyuan Yao, Zhenhua Song, Yuyang Zhou, Tenglong Ao, Baoquan Chen, and Libin Liu. Moconvq: Unified physics- based motion control via scalable discrete representations. In ACM SIGGRAPH 2024 Journal Proceedings, 2023

    2024

  62. [70]

    Neu- ral3points: Learning to generate physically realistic full-body motion for virtual reality users

    Yongjing Ye, Libin Liu, Lei Hu, and Shihong Xia. Neu- ral3points: Learning to generate physically realistic full-body motion for virtual reality users. In Computer Graphics Forum. Wiley Online Library, 2022. 2

    2022

  63. [71]

    Physdiff: Physics-guided human motion diffusion model

    Ye Yuan, Jiaming Song, Umar Iqbal, Arash Vahdat, and Jan Kautz. Physdiff: Physics-guided human motion diffusion model. In Proceedings of the IEEE/CVF International Con- ference on Computer Vision, pages 16010–16021, 2023. 2

    2023

  64. [72]

    Mode-adaptive neural networks for quadruped motion control

    He Zhang, Sebastian Starke, Taku Komura, and Jun Saito. Mode-adaptive neural networks for quadruped motion control. ACM Trans. Graph., 2018. 2, 6, 7

    2018

  65. [73]

    Adding conditional control to text-to-image diffusion models

    Lvmin Zhang, Anyi Rao, and Maneesh Agrawala. Adding conditional control to text-to-image diffusion models. In Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV), pages 3836–3847, 2023. 2

    2023

  66. [74]

    Motiondiffuse: Text- driven human motion generation with diffusion model

    Mingyuan Zhang, Zhongang Cai, Liang Pan, Fangzhou Hong, Xinying Guo, Lei Yang, and Ziwei Liu. Motiondiffuse: Text- driven human motion generation with diffusion model. IEEE Transactions on Pattern Analysis & Machine Intelligence ,

  67. [75]

    Dart: A diffusion- based autoregressive motion model for real-time text-driven motion control, 2024

    Kaifeng Zhao, Gen Li, and Siyu Tang. Dart: A diffusion- based autoregressive motion model for real-time text-driven motion control, 2024. 1

    2024

  68. [76]

    M. Zhao, M. Liu, B. Ren, and S. Dai. Modiff: Action- conditioned 3d motion generation with denoising diffusion probabilistic models. arXiv preprint arXiv:2301.03949, 2023. 2

    2023 arXiv

  69. [77]

    M. Zhao, M. Liu, B. Ren, and S. Dai. Denoising diffusion probabilistic models for action-conditioned 3d motion genera- tion. In ICASSP 2024 - 2024 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 2024. 2

    2024

  70. [78]

    Media2face: Co-speech facial animation generation with multi-modality guidance

    Qingcheng Zhao, Pengyu Long, Qixuan Zhang, Dafei Qin, Han Liang, Longwen Zhang, Yingliang Zhang, Jingyi Yu, and Lan Xu. Media2face: Co-speech facial animation generation with multi-modality guidance. In ACM SIGGRAPH 2024 Conference Papers, 2024. 2

    2024

  71. [79]

    Emdm: Efficient motion dif- fusion model for fast, high-quality motion generation

    Wenyang Zhou, Zhiyang Dou, Zeyu Cao, Zhouyingcheng Liao, Jingbo Wang, Wenjia Wang, Yuan Liu, Taku Komura, Wenping Wang, and Lingjie Liu. Emdm: Efficient motion dif- fusion model for fast, high-quality motion generation. ECCV,

  72. [80]

    On the continuity of rotation representations in neural networks

    Yi Zhou, Connelly Barnes, Jingwan Lu, Jimei Yang, and Hao Li. On the continuity of rotation representations in neural networks. 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2018. 3

    2019

  73. [81]

    Ude: A unified driv- ing engine for human motion generation

    Zixiang Zhou and Baoyuan Wang. Ude: A unified driv- ing engine for human motion generation. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 5632–5641, 2023. 2 11

    2023

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.