pith. sign in

arxiv: 2606.21135 · v1 · pith:4KHXAPHSnew · submitted 2026-06-19 · 💻 cs.CV · cs.GR· cs.RO

Odoriko: A Shape-Aware Multimodal Diffusion Framework for Human Motion

Dongseok Shim , Julian Tanke , Kengo Uchida , Christian Simon , Koichi Saito , Takashi Shibuya , Shusuke Takahashi , Yuki Mitsufuji This is my paper

Pith reviewed 2026-06-26 14:37 UTC · model grok-4.3

classification 💻 cs.CV cs.GRcs.RO
keywords human motion generationmultimodal diffusionshape-aware generationtext-to-motionmusic-to-dancevideo-to-motionmorphology estimation
0
0 comments X

The pith

A single diffusion model generates motion consistent with the mover's body shape and gender from text, music, or video.

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

The paper introduces Odoriko as the first unified multimodal framework for human motion generation that directly incorporates subject bio-morphological information such as gender and body shape. It produces motion outputs consistent with the specific person performing the action rather than averaging across subjects, and it works across text, music, and video inputs in one model. When morphological details are missing, the same framework recovers them alongside the motion. A sympathetic reader would care because prior unified models treat all bodies as equivalent, which limits realism in generated movements. Experiments indicate the model matches or exceeds specialized single-modality approaches on standard benchmarks while adding this new capability.

Core claim

Odoriko is the first unified multimodal motion generation framework that reflects subject bio-morphological information directly in synthesized motion output. Rather than averaging over subject variation, Odoriko generates motion that is consistent with who is moving, not just what they are asked to do, across text, music, and video conditions within a single model. When explicit morphological information is unavailable, Odoriko additionally recovers subject morphology alongside motion, unifying estimation and generation in one framework.

What carries the argument

Shape-aware multimodal diffusion model that conditions motion synthesis on bio-morphological factors such as gender and body shape.

If this is right

  • Matches or exceeds prior specialized models on text-to-motion, music-to-dance, and video-to-motion benchmarks.
  • Enables morphology-consistent generation that no existing unified framework supports.
  • Unifies morphology estimation and motion generation within one model.
  • Produces motion outputs that reflect who is moving rather than averaging over subjects.

Where Pith is reading between the lines

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

  • Applications in animation and virtual reality could use the model to create personalized avatars without building separate body-type networks.
  • The approach may extend to real-time systems where body measurements are estimated from a single video frame.
  • Testing on body shapes far outside the training distribution would reveal how far the conditioning generalizes.

Load-bearing premise

Morphological factors such as gender and body shape produce distinct kinematic signatures that can be directly reflected in a unified diffusion model without requiring separate per-modality or per-body-type architectures.

What would settle it

Generate motions for the same prompt but with different body shapes and check whether measurable kinematic differences appear in joint trajectories or velocities that align with real captured data for those shapes.

Figures

Figures reproduced from arXiv: 2606.21135 by Christian Simon, Dongseok Shim, Julian Tanke, Kengo Uchida, Koichi Saito, Shusuke Takahashi, Takashi Shibuya, Yuki Mitsufuji.

Figure 1
Figure 1. Figure 1: We present Odoriko, a shape-aware multimodal human motion framework that unifies motion generation and 3D pose estimation within a single model. Our approach explicitly models subject-specific body shape, enabling both shape-conditioned motion synthesis and shape-aware pose estimation under diverse multimodal inputs. fragmented design requires separate architectures and training pipelines for each modality… view at source ↗
Figure 2
Figure 2. Figure 2: Odoriko visualizations showing shape-dependent motion variations under iden￾tical text or music conditions. motion estimation from videos and 2D keypoints within a single shared archi￾tecture, enabling controllable and anatomically coherent motion modeling across tasks and modalities. We represent subject shape using gender and SMPL [40] β parameters, pro￾viding a compact yet expressive encoding of body mo… view at source ↗
Figure 3
Figure 3. Figure 3: Odoriko Overview. Temporally aligned multimodal inputs are fused with noisy motion tokens, while their global features condition the denoising timestep t. A shape token constructed from gender g and body shape β is integrated via joint attention, enabling explicit morphology-aware generation. The model outputs denoised motion xˆ0, and additionally predicts shape parameters in shape estimation mode. For cla… view at source ↗
read the original abstract

Human motion generation has been widely studied across diverse input modalities, text, music, and video, and recent efforts have unified these into single multimodal frameworks. However, while morphological factors such as gender and body shape are known to produce distinct kinematic signatures, no existing unified framework incorporates this into generation, treating all subjects as morphologically equivalent. We present Odoriko, the first unified multimodal motion generation framework that reflects subject bio-morphological information directly in synthesized motion output. Rather than averaging over subject variation, Odoriko generates motion that is consistent with who is moving, not just what they are asked to do, across text, music, and video conditions within a single model. When explicit morphological information is unavailable, Odoriko additionally recovers subject morphology alongside motion, unifying estimation and generation in one framework. Extensive experiments across text-to-motion, music-to-dance, and video-to-motion benchmarks demonstrate that Odoriko matches or exceeds prior specialized models on standard metrics, while enabling morphology-consistent generation that no existing unified framework supports.

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

1 major / 0 minor

Summary. The manuscript introduces Odoriko, the first unified multimodal diffusion framework for human motion generation that directly incorporates subject bio-morphological parameters (gender, body shape) to synthesize motion consistent with the individual across text, music, and video conditioning within a single model. It further unifies generation with morphology estimation when explicit shape information is unavailable. The central claim is that this approach produces subject-specific kinematic signatures rather than averaging over morphological variation, and that experiments on standard text-to-motion, music-to-dance, and video-to-motion benchmarks show performance matching or exceeding prior specialized models while adding morphology consistency not supported by existing unified frameworks.

Significance. If the empirical results hold, the work would address a recognized gap in multimodal motion synthesis by moving beyond morphologically agnostic models. Incorporating biomechanical factors into a single diffusion backbone could improve realism for personalized animation and interaction tasks. The joint estimation-generation capability is a practical addition that aligns with established observations that morphology influences kinematics.

major comments (1)
  1. [Abstract] Abstract: the claim that Odoriko 'matches or exceeds prior specialized models on standard metrics' is presented without any numerical values, tables, ablation studies, or error bars. This absence prevents evaluation of whether the central empirical claim is supported and whether morphology-consistent generation comes at any cost to standard metrics.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for their review and for highlighting the need for clearer empirical support in the abstract. We address the single major comment below and will revise the manuscript accordingly.

read point-by-point responses

  1. Referee: [Abstract] Abstract: the claim that Odoriko 'matches or exceeds prior specialized models on standard metrics' is presented without any numerical values, tables, ablation studies, or error bars. This absence prevents evaluation of whether the central empirical claim is supported and whether morphology-consistent generation comes at any cost to standard metrics.

    Authors: We agree that the abstract, as written, states the performance claim without supporting numbers. The full manuscript (Sections 4–5 and Tables 1–4) contains the requested quantitative comparisons, including FID, R-Precision, Diversity, and Multimodality scores on HumanML3D, AIST++, and other benchmarks, plus ablations isolating the morphology conditioning and error bars from multiple runs. These results show Odoriko matching or exceeding specialized baselines while adding morphology consistency. To directly address the concern, we will revise the abstract to include 2–3 key numerical highlights (e.g., average FID improvement and morphology consistency metric) and reference the tables. revision: yes

Circularity Check

0 steps flagged

No significant circularity

full rationale

The paper presents an empirical multimodal diffusion model for human motion generation that conditions on morphological parameters. No derivation chain, equations, or first-principles predictions are described in the abstract or provided text that reduce to inputs by construction, fitted parameters renamed as predictions, or self-citation load-bearing arguments. The central claim rests on model architecture and benchmark evaluations rather than any self-referential mathematical reduction, making the work self-contained against external benchmarks.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Only the abstract is available; no free parameters, axioms or invented entities can be extracted or verified from the provided text.

pith-pipeline@v0.9.1-grok · 5737 in / 1090 out tokens · 13330 ms · 2026-06-26T14:37:22.282056+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

77 extracted references · 7 canonical work pages · 4 internal anchors

  1. [1]

    In: European Conference on Computer Vision

    Árbol, B.R., Casas, D.: Bodyshapegpt: Smpl body shape manipulation with llms. In: European Conference on Computer Vision. pp. 388–396. Springer (2024) 2, 4

    2024

  2. [2]

    Perception & psychophysics23(2), 145–152 (1978) 2, 4, 24

    Barclay, C.D., Cutting, J.E., Kozlowski, L.T.: Temporal and spatial factors in gait perception that influence gender recognition. Perception & psychophysics23(2), 145–152 (1978) 2, 4, 24

    1978

  3. [3]

    In: Proceedings of the AAAI Conference on Artificial Intelligence

    Bian, Y., Zeng, A., Ju, X., Liu, X., Zhang, Z., Liu, W., Xu, Q.: Motioncraft: Craft- ing whole-body motion with plug-and-play multimodal controls. In: Proceedings of the AAAI Conference on Artificial Intelligence. vol. 39, pp. 1880–1888 (2025) 4, 10, 11

    2025

  4. [4]

    IEEE transactions on pattern analysis and machine intelligence43(1), 172–186 (2019) 7

    Cao, Z., Hidalgo, G., Simon, T., Wei, S.E., Sheikh, Y.: Openpose: Realtime multi- person 2d pose estimation using part affinity fields. IEEE transactions on pattern analysis and machine intelligence43(1), 172–186 (2019) 7

    2019

  5. [5]

    arXiv preprint arXiv:2502.02063 (2025) 25

    Chang, C.J., Liu, Q.T., Zhou, H., Pavlovic, V., Kapadia, M.: Casim: Com- posite aware semantic injection for text to motion generation. arXiv preprint arXiv:2502.02063 (2025) 25

    work page arXiv 2025

  6. [6]

    In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition

    Chen, X., Jiang, B., Liu, W., Huang, Z., Fu, B., Chen, T., Yu, G.: Executing your commands via motion diffusion in latent space. In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition. pp. 18000–18010 (2023) 4, 10

    2023

  7. [7]

    In: Proceedings of the Computer Vision and Pattern Recognition Confer- ence

    Cheng, H.K., Ishii, M., Hayakawa, A., Shibuya, T., Schwing, A., Mitsufuji, Y.: Mmaudio: Taming multimodal joint training for high-quality video-to-audio syn- thesis. In: Proceedings of the Computer Vision and Pattern Recognition Confer- ence. pp. 28901–28911 (2025) 7, 25

    2025

  8. [8]

    Choutas, V., Müller, L., Huang, C.H.P., Tang, S., Tzionas, D., Black, M.J.: Accu- rate3dbodyshaperegressionusingmetricandsemanticattributes.In:Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. pp. 2718–2728 (2022) 2, 4, 13

    2022

  9. [9]

    Perception7(4), 393–405 (1978) 2, 4

    Cutting, J.E.: Generation of synthetic male and female walkers through manipu- lation of a biomechanical invariant. Perception7(4), 393–405 (1978) 2, 4

    1978

  10. [10]

    In: European Confer- ence on Computer Vision (2024) 1, 4, 10, 11

    Dai, W., Chen, L.H., Wang, J., Liu, J., Dai, B., Tang, Y.: Motionlcm: Real-time controllable motion generation via latent consistency model. In: European Confer- ence on Computer Vision (2024) 1, 4, 10, 11

    2024

  11. [11]

    Jukebox: A Generative Model for Music

    Dhariwal, P., Jun, H., Payne, C., Kim, J.W., Radford, A., Sutskever, I.: Jukebox: A generative model for music. arXiv preprint arXiv:2005.00341 (2020) 7

    work page internal anchor Pith review Pith/arXiv arXiv 2005

  12. [12]

    Advances in neural information processing systems34, 8780–8794 (2021) 21

    Dhariwal, P., Nichol, A.: Diffusion models beat gans on image synthesis. Advances in neural information processing systems34, 8780–8794 (2021) 21

    2021

  13. [13]

    In: Forty-first international conference on machine learning (2024) 7, 21, 25

    Esser, P., Kulal, S., Blattmann, A., Entezari, R., Müller, J., Saini, H., Levi, Y., Lorenz, D., Sauer, A., Boesel, F., et al.: Scaling rectified flow transformers for high-resolution image synthesis. In: Forty-first international conference on machine learning (2024) 7, 21, 25

    2024

  14. [14]

    In: 16 D

    Gong, K., Lian, D., Chang, H., Guo, C., Jiang, Z., Zuo, X., Mi, M.B., Wang, X.: Tm2d: Bimodality driven 3d dance generation via music-text integration. In: 16 D. Shimet al. Proceedings of the IEEE/CVF International Conference on Computer Vision. pp. 9942–9952 (2023) 10

    2023

  15. [15]

    In: Proceedings of the IEEE/CVF International Confer- ence on Computer Vision

    Gralnik, O., Gafni, G., Shamir, A.: Semantify: Simplifying the control of 3d mor- phable models using clip. In: Proceedings of the IEEE/CVF International Confer- ence on Computer Vision. pp. 14554–14564 (2023) 4

    2023

  16. [16]

    In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition

    Guo, C., Mu, Y., Javed, M.G., Wang, S., Cheng, L.: Momask: Generative masked modeling of 3d human motions. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. pp. 1900–1910 (2024) 1, 4, 5, 10, 11, 12, 13, 22

    1900

  17. [17]

    In: Proceedings of the IEEE/CVF con- ference on computer vision and pattern recognition

    Guo, C., Zou, S., Zuo, X., Wang, S., Ji, W., Li, X., Cheng, L.: Generating diverse and natural 3d human motions from text. In: Proceedings of the IEEE/CVF con- ference on computer vision and pattern recognition. pp. 5152–5161 (2022) 5, 9, 10, 11, 20, 23, 26, 27

    2022

  18. [18]

    In: Proceedings of the 28th ACM international conference on multimedia

    Guo, C., Zuo, X., Wang, S., Zou, S., Sun, Q., Deng, A., Gong, M., Cheng, L.: Action2motion: Conditioned generation of 3d human motions. In: Proceedings of the 28th ACM international conference on multimedia. pp. 2021–2029 (2020) 9

    2021

  19. [19]

    In: Proceedings of the IEEE/CVF International Conference on Computer Vision

    Guo, Z., Hu, Z., Soh, D.W., Zhao, N.: Motionlab: Unified human motion gener- ation and editing via the motion-condition-motion paradigm. In: Proceedings of the IEEE/CVF International Conference on Computer Vision. pp. 13869–13879 (2025) 4

    2025

  20. [20]

    Advances in neural information processing systems33, 6840–6851 (2020) 6, 9, 25

    Ho, J., Jain, A., Abbeel, P.: Denoising diffusion probabilistic models. Advances in neural information processing systems33, 6840–6851 (2020) 6, 9, 25

    2020

  21. [21]

    In: NeurIPS 2021 Workshop on Deep Generative Models and Downstream Applications (2021) 21

    Ho, J., Salimans, T.: Classifier-free diffusion guidance. In: NeurIPS 2021 Workshop on Deep Generative Models and Downstream Applications (2021) 21

    2021

  22. [22]

    ACM Transactions on Graphics (TOG)36(4), 1–13 (2017) 1

    Holden, D., Komura, T., Saito, J.: Phase-functioned neural networks for character control. ACM Transactions on Graphics (TOG)36(4), 1–13 (2017) 1

    2017

  23. [23]

    ACM Transactions on Graphics (2016) 6

    Holden, D., Saito, J., Komura, T.: A deep learning framework for character motion synthesis and editing. ACM Transactions on Graphics (2016) 6

    2016

  24. [24]

    Archives of physical medicine and rehabilitation95(10), 1860–1869 (2014) 2, 4, 13, 24

    Hsue, B.J., Su, F.C.: Effects of age and gender on dynamic stability during stair descent. Archives of physical medicine and rehabilitation95(10), 1860–1869 (2014) 2, 4, 13, 24

    2014

  25. [25]

    6m: Large scale datasets and predictive methods for 3d human sensing in natural environments

    Ionescu, C., Papava, D., Olaru, V., Sminchisescu, C.: Human3. 6m: Large scale datasets and predictive methods for 3d human sensing in natural environments. IEEE transactions on pattern analysis and machine intelligence36(7), 1325–1339 (2013) 10, 27

    2013

  26. [26]

    In: Proceedings of the IEEE/CVF International Conference on Computer Vision

    Kaufmann, M., Song, J., Guo, C., Shen, K., Jiang, T., Tang, C., Zárate, J.J., Hilliges, O.: Emdb: The electromagnetic database of global 3d human pose and shape in the wild. In: Proceedings of the IEEE/CVF International Conference on Computer Vision. pp. 14632–14643 (2023) 10, 22, 27

    2023

  27. [27]

    Kim, J., Oh, H., Kim, S., Tong, H., Lee, S.: A brand new dance partner: Music- conditionedpluralisticdancingcontrolledbymultipledancegenres.In:Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. pp. 3490–3500 (2022) 1, 11

    2022

  28. [28]

    Labs,B.F.,Batifol,S.,Blattmann,A.,Boesel,F.,Consul,S.,Diagne,C.,Dockhorn, T., English, J., English, Z., Esser, P., Kulal, S., Lacey, K., Levi, Y., Li, C., Lorenz, D., Müller, J., Podell, D., Rombach, R., Saini, H., Sauer, A., Smith, L.: Flux.1 kontext: Flow matching for in-context image generation and editing in latent space (2025),https://arxiv.org/abs/2...

    work page internal anchor Pith review Pith/arXiv arXiv 2025

  29. [29]

    In: In- ternational Conference on Machine Learning

    Lee, W., Jeong, J., Moon, T., Kim, H.J., Kim, J., Kim, G., Lee, B.U.: How to move your dragon: Text-to-motion synthesis for large-vocabulary objects. In: In- ternational Conference on Machine Learning. pp. 33110–33128. PMLR (2025) 20 Odoriko 17

    2025

  30. [30]

    ACM Transactions on Graphics (TOG)42(6), 1–11 (2023) 1

    Li, J., Wu, J., Liu, C.K.: Object motion guided human motion synthesis. ACM Transactions on Graphics (TOG)42(6), 1–11 (2023) 1

    2023

  31. [31]

    In: International Conference on Computer Vision (2025) 4, 6, 10, 11, 12, 21, 26

    Li, J., Cao, J., Zhang, H., Rempe, D., Kautz, J., Iqbal, U., Yuan, Y.: Genmo: A generalist model for human motion. In: International Conference on Computer Vision (2025) 4, 6, 10, 11, 12, 21, 26

    2025

  32. [32]

    arXiv preprint arXiv:2410.20389 (2024) 1, 6, 11

    Li, R., Zhang, H., Zhang, Y., Zhang, Y., Zhang, Y., Guo, J., Zhang, Y., Li, X., Liu,Y.:Lodge++:High-qualityandlongdancegenerationwithvividchoreography patterns. arXiv preprint arXiv:2410.20389 (2024) 1, 6, 11

    work page arXiv 2024

  33. [33]

    In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition

    Li, R., Zhang, Y., Zhang, Y., Zhang, H., Guo, J., Zhang, Y., Liu, Y., Li, X.: Lodge: A coarse to fine diffusion network for long dance generation guided by the characteristic dance primitives. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. pp. 1524–1534 (2024) 1, 4, 6, 10, 11, 22

    2024

  34. [34]

    In: Proceed- ings of the IEEE/CVF International Conference on Computer Vision

    Li, R., Zhao, J., Zhang, Y., Su, M., Ren, Z., Zhang, H., Tang, Y., Li, X.: Finedance: A fine-grained choreography dataset for 3d full body dance generation. In: Proceed- ings of the IEEE/CVF International Conference on Computer Vision. pp. 10234– 10243 (2023) 4, 8, 10, 22, 27

    2023

  35. [35]

    In: Proceedings of the IEEE/CVF international conference on computer vision

    Li, R., Yang, S., Ross, D.A., Kanazawa, A.: Ai choreographer: Music conditioned 3d dance generation with aist++. In: Proceedings of the IEEE/CVF international conference on computer vision. pp. 13401–13412 (2021) 1, 4, 10, 11, 22, 27

    2021

  36. [36]

    Liao, T.H., Zhou, Y., Shen, Y., Huang, C.H.P., Mitra, S., Huang, J.B., Bhat- tacharya, U.: Shape my moves: Text-driven shape-aware synthesis of human mo- tions.In:ProceedingsoftheComputerVisionandPatternRecognitionConference. pp. 1917–1928 (2025) 2, 4, 13, 20

    1917

  37. [37]

    arXiv preprint arXiv:2411.17335 (2024) 11

    Ling, Z., Han, B., Li, S., Cheng, J., Shen, H., Zou, C.: Versatilemotion: A unified framework for motion synthesis and comprehension. arXiv preprint arXiv:2411.17335 (2024) 11

    work page arXiv 2024

  38. [38]

    In: Proceedings of the Thirty-Third In- ternational Joint Conference on Artificial Intelligence, IJCAI-24

    Ling, Z., Han, B., Wong, Y., Lin, H., Kankanhalli, M., Geng, W.: Mcm: Multi- condition motion synthesis framework. In: Proceedings of the Thirty-Third In- ternational Joint Conference on Artificial Intelligence, IJCAI-24. pp. 1083–1091 (2024) 4, 10, 11

    2024

  39. [39]

    Flow Matching for Generative Modeling

    Lipman, Y., Chen, R.T., Ben-Hamu, H., Nickel, M., Le, M.: Flow matching for generative modeling. arXiv preprint arXiv:2210.02747 (2022) 25

    work page internal anchor Pith review Pith/arXiv arXiv 2022

  40. [40]

    ACM Transactions on Graphics (TOG)34(6), 1–16 (2015) 3, 5, 20

    Loper, M., Mahmood, N., Romero, J., Pons-Moll, G., Black, M.J.: Smpl: a skinned multi-person linear model. ACM Transactions on Graphics (TOG)34(6), 1–16 (2015) 3, 5, 20

    2015

  41. [41]

    In: International Conference on Learning Representations (2019) 21

    Loshchilov, I., Hutter, F.: Decoupled weight decay regularization. In: International Conference on Learning Representations (2019) 21

    2019

  42. [42]

    Advances in Neural Information Processing Systems37, 28051–28077 (2024) 4, 11

    Luo, M., Hou, R., Li, Z., Chang, H., Liu, Z., Wang, Y., Shan, S.: M3 gpt: An advanced multimodal, multitask framework for motion comprehension and genera- tion. Advances in Neural Information Processing Systems37, 28051–28077 (2024) 4, 11

    2024

  43. [43]

    In: Proceedings of the IEEE/CVF international conference on computer vision

    Mahmood, N., Ghorbani, N., Troje, N.F., Pons-Moll, G., Black, M.J.: Amass: Archive of motion capture as surface shapes. In: Proceedings of the IEEE/CVF international conference on computer vision. pp. 5442–5451 (2019) 9

    2019

  44. [44]

    In: Proceedings of the Computer Vision and Pattern Recognition Conference

    Maldonado, G., Pazho, A.D., Noghre, G.A., Katariya, V., Tabkhi, H.: Moclip motion-aware fine-tuning and distillation of clip for human motion generation. In: Proceedings of the Computer Vision and Pattern Recognition Conference. pp. 2931–2941 (2025) 25 18 D. Shimet al

    2025

  45. [45]

    In: Proceedings of the Computer Vision and Pattern Recognition Conference

    Meng, Z., Xie, Y., Peng, X., Han, Z., Jiang, H.: Rethinking diffusion for text-driven human motion generation: Redundant representations, evaluation, and masked au- toregression. In: Proceedings of the Computer Vision and Pattern Recognition Conference. pp. 27859–27871 (2025) 4, 5, 23

    2025

  46. [46]

    Newnes (2012) 1

    Parent, R.: Computer animation: algorithms and techniques. Newnes (2012) 1

    2012

  47. [47]

    io/(2021), project page 26

    Park, K., Sinha, U., Barron, J.T., Bouaziz, S., Goldman, D.B., Seitz, S.M., Martin- Brualla, R.: Nerfies: Deformable neural radiance fields.https://nerfies.github. io/(2021), project page 26

    2021

  48. [48]

    In: Proceedings oftheIEEE/CVFConferenceonComputerVisionandPatternRecog- nition

    Petrovich, M., Litany, O., Iqbal, U., Black, M.J., Varol, G., Bin Peng, X., Rempe, D.: Multi-track timeline control for text-driven 3d human motion generation. In: Proceedings oftheIEEE/CVFConferenceonComputerVisionandPatternRecog- nition. pp. 1911–1921 (2024) 6

    1911

  49. [49]

    In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition

    Pinyoanuntapong, E., Wang, P., Lee, M., Chen, C.: Mmm: Generative masked motion model. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. pp. 1546–1555 (2024) 1, 4, 10, 11

    2024

  50. [50]

    In: International conference on machine learning

    Radford, A., Kim, J.W., Hallacy, C., Ramesh, A., Goh, G., Agarwal, S., Sastry, G., Askell, A., Mishkin, P., Clark, J., et al.: Learning transferable visual models from natural language supervision. In: International conference on machine learning. pp. 8748–8763. PmLR (2021) 7, 20, 25

    2021

  51. [51]

    Journal of machine learning research21(140), 1–67 (2020) 7, 20

    Raffel, C., Shazeer, N., Roberts, A., Lee, K., Narang, S., Matena, M., Zhou, Y., Li, W., Liu, P.J.: Exploring the limits of transfer learning with a unified text-to-text transformer. Journal of machine learning research21(140), 1–67 (2020) 7, 20

    2020

  52. [52]

    In: European Conference on Computer Vision

    Ren, Z., Huang, S., Li, X.: Realistic human motion generation with cross-diffusion models. In: European Conference on Computer Vision. pp. 345–362. Springer (2024) 6

    2024

  53. [53]

    In: SIGGRAPH Asia 2024 Conference Papers

    Shen,Z.,Pi,H.,Xia,Y.,Cen,Z.,Peng,S.,Hu,Z.,Bao,H.,Hu,R.,Zhou,X.:World- grounded human motion recovery via gravity-view coordinates. In: SIGGRAPH Asia 2024 Conference Papers. pp. 1–11 (2024) 1, 4, 11, 12

    2024

  54. [54]

    In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition

    Shin, S., Kim, J., Halilaj, E., Black, M.J.: Wham: Reconstructing world-grounded humans with accurate 3d motion. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. pp. 2070–2080 (2024) 1, 4, 11, 12

    2070

  55. [55]

    In: Proceed- ings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition

    Siyao, L., Yu, W., Gu, T., Lin, C., Wang, Q., Qian, C., Loy, C.C., Liu, Z.: Bailando: 3d dance generation by actor-critic gpt with choreographic memory. In: Proceed- ings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. pp. 11050–11059 (2022) 1, 4, 11

    2022

  56. [56]

    Song,J.,Meng,C.,Ermon,S.:Denoisingdiffusionimplicitmodels.In:International Conference on Learning Representations (2021) 9, 14

    2021

  57. [57]

    ACM Transactions on Graphics (TOG)40(4), 1–16 (2021) 1

    Starke, S., Zhao, Y., Zinno, F., Komura, T.: Neural animation layering for syn- thesizing martial arts movements. ACM Transactions on Graphics (TOG)40(4), 1–16 (2021) 1

    2021

  58. [58]

    In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition

    Sun, Y., Bao, Q., Liu, W., Mei, T., Black, M.J.: Trace: 5d temporal regression of avatars with dynamic cameras in 3d environments. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. pp. 8856– 8866 (2023) 1, 10, 11, 12

    2023

  59. [59]

    In: The Eleventh International Conference on Learning Representations (2023) 1, 4, 6, 10, 11, 12, 13, 20, 22

    Tevet, G., Raab, S., Gordon, B., Shafir, Y., Cohen-Or, D., Bermano, A.H.: Human motion diffusion model. In: The Eleventh International Conference on Learning Representations (2023) 1, 4, 6, 10, 11, 12, 13, 20, 22

    2023

  60. [60]

    In: European Conference on Computer Vision

    Tripathi, S., Taheri, O., Lassner, C., Black, M., Holden, D., Stoll, C.: Humos: Human motion model conditioned on body shape. In: European Conference on Computer Vision. pp. 133–152. Springer (2024) 4 Odoriko 19

    2024

  61. [61]

    Journal of vision2(5), 2–2 (2002) 2, 4, 13, 24

    Troje, N.F.: Decomposing biological motion: A framework for analysis and synthe- sis of human gait patterns. Journal of vision2(5), 2–2 (2002) 2, 4, 13, 24

    2002

  62. [62]

    In: Proceedings of the IEEE/CVF conference on computer vision and pattern recog- nition

    Tseng, J., Castellon, R., Liu, K.: Edge: Editable dance generation from music. In: Proceedings of the IEEE/CVF conference on computer vision and pattern recog- nition. pp. 448–458 (2023) 4, 7, 11, 20

    2023

  63. [63]

    In: Proceedings of the Computer Vision and Pattern Recog- nition Conference

    Uchida, K., Shibuya, T., Takida, Y., Murata, N., Tanke, J., Takahashi, S., Mit- sufuji, Y.: Mola: Motion generation and editing with latent diffusion enhanced by adversarial training. In: Proceedings of the Computer Vision and Pattern Recog- nition Conference. pp. 2910–2919 (2025) 1, 10, 11, 12, 13, 22

    2025

  64. [64]

    Advances in neural information pro- cessing systems30(2017) 20

    Vaswani,A.,Shazeer,N.,Parmar,N.,Uszkoreit,J.,Jones,L.,Gomez,A.N.,Kaiser, Ł., Polosukhin, I.: Attention is all you need. Advances in neural information pro- cessing systems30(2017) 20

    2017

  65. [65]

    In: Proceedings of the European conference on computer vision (ECCV)

    Von Marcard, T., Henschel, R., Black, M.J., Rosenhahn, B., Pons-Moll, G.: Re- covering accurate 3d human pose in the wild using imus and a moving camera. In: Proceedings of the European conference on computer vision (ECCV). pp. 601–617 (2018) 10, 22, 27

    2018

  66. [66]

    Wan: Open and Advanced Large-Scale Video Generative Models

    Wan, T., Wang, A., Ai, B., Wen, B., Mao, C., Xie, C.W., Chen, D., Yu, F., Zhao, H., Yang, J., et al.: Wan: Open and advanced large-scale video generative models. arXiv preprint arXiv:2503.20314 (2025) 25

    work page internal anchor Pith review Pith/arXiv arXiv 2025

  67. [67]

    In: Proceedings of the AAAI Conference on Artificial Intelligence (2026) 2, 4

    Wang, X., Xu, K., Li, F., Sheng, C., Yu, J., Mu, Y.: Generating attribute-aware human motions from textual prompt. In: Proceedings of the AAAI Conference on Artificial Intelligence (2026) 2, 4

    2026

  68. [68]

    In: European Conference on Computer Vision

    Wang, Y., Wang, Z., Liu, L., Daniilidis, K.: Tram: Global trajectory and motion of 3d humans from in-the-wild videos. In: European Conference on Computer Vision. pp. 467–487. Springer (2024) 1, 4, 7, 10, 11, 12, 20

    2024

  69. [69]

    In: 2013 IEEE International Confer- ence on Robotics and Automation

    Yamane, K., Revfi, M., Asfour, T.: Synthesizing object receiving motions of hu- manoid robots with human motion database. In: 2013 IEEE International Confer- ence on Robotics and Automation. pp. 1629–1636. IEEE (2013) 1

    2013

  70. [70]

    In: Proceedings of the IEEE/CVF International Conference on Computer Vision

    Yang, Z., Zeng, A., Yuan, C., Li, Y.: Effective whole-body pose estimation with two-stages distillation. In: Proceedings of the IEEE/CVF International Conference on Computer Vision. pp. 4210–4220 (2023) 7, 10

    2023

  71. [71]

    In: Proceedings of theIEEE/CVF conference on computer vision and pattern recognition

    Zhang, J., Zhang, Y., Cun, X., Zhang, Y., Zhao, H., Lu, H., Shen, X., Shan, Y.: Generating human motion from textual descriptions with discrete representa- tions. In: Proceedings of theIEEE/CVF conference on computer vision and pattern recognition. pp. 14730–14740 (2023) 1, 10, 12, 13

    2023

  72. [72]

    IEEE transactions on pattern analysis and machine intelligence46(6), 4115–4128 (2024) 4, 10, 11

    Zhang, M., Cai, Z., Pan, L., Hong, F., Guo, X., Yang, L., Liu, Z.: Motiondiffuse: Text-driven human motion generation with diffusion model. IEEE transactions on pattern analysis and machine intelligence46(6), 4115–4128 (2024) 4, 10, 11

    2024

  73. [73]

    In: Proceedings of the IEEE/CVF International Conference on Computer Vision

    Zhang, M., Guo, X., Pan, L., Cai, Z., Hong, F., Li, H., Yang, L., Liu, Z.: Re- modiffuse: Retrieval-augmented motion diffusion model. In: Proceedings of the IEEE/CVF International Conference on Computer Vision. pp. 364–373 (2023) 6

    2023

  74. [74]

    In: European Conference on Computer Vision

    Zhang, M., Jin, D., Gu, C., Hong, F., Cai, Z., Huang, J., Zhang, C., Guo, X., Yang, L., He, Y., et al.: Large motion model for unified multi-modal motion generation. In: European Conference on Computer Vision. pp. 397–421. Springer (2024) 10

    2024

  75. [75]

    Advances in Neural Information Processing Systems36, 49842–49869 (2023) 9, 14

    Zhao, W., Bai, L., Rao, Y., Zhou, J., Lu, J.: Unipc: A unified predictor-corrector framework for fast sampling of diffusion models. Advances in Neural Information Processing Systems36, 49842–49869 (2023) 9, 14

    2023

  76. [76]

    In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition

    Zhou, Y., Barnes, C., Lu, J., Yang, J., Li, H.: On the continuity of rotation rep- resentations in neural networks. In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition. pp. 5745–5753 (2019) 6, 20 20 D. Shimet al. A Implementation Details A.1 Network Details T able 8:Transformer architecture details of Odoriko. Latent dim ...

    2019

  77. [77]

    test split using both kinematic and geometric motion features. Specifically, we compute FID and Diversity (Div) in two feature spaces: (i) kinematic fea- tures, denoted as FIDk and Divk, which capture physical motion dynamics, and (ii) geometric features, denoted as FIDg and Divg, which reflect overall dance choreography. Both feature representations are ...