Pith. sign in

REVIEW 4 major objections 6 minor 40 references

CASIM: Composite Aware Semantic Injection for Text to Motion Generation

T0 review · 4 major / 6 minor · reviewed 2026-08-09 · deepseek-v4-flash

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

Pith's one-line read Replacing fixed-length CLIP text embeddings with token-level embeddings plus a learned text-motion aligner improves motion quality, alignment, and retrieval across five text-to-motion models.

desk verdict CASIM's token-level injection idea is worth taking seriously, but the main comparison tables mix architectures and baselines in ways that inflate the reported gains. read the letter →

arxiv 2502.02063 v1 pith:PORDPCX5 submitted 2025-02-04 cs.CV cs.AIcs.GR

Che-Jui Chang , Qingze Tony Liu , Honglu Zhou , Vladimir Pavlovic , Mubbasir Kapadia This is my paper · ORCID
classification cs.CVcs.AIcs.GR
keywords text-to-motiongenerationsemanticinjectiontoken-leveltextembeddingstext-motionalignmentmulti-headattentiondiffusionmodelsautoregressivehumanmotion
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

This paper argues that the standard practice of compressing a text prompt into one fixed-length CLIP [CLS] embedding loses the composite, order-sensitive structure of human motion descriptions, and that this loss limits both the quality and the controllability of generated motion. The proposed CASIM mechanism instead keeps the per-token text embeddings and adds a learned text-motion aligner that lets each motion frame attend to the words most relevant to it. Across five state-of-the-art generators—MDM, MotionDiffuse, T2MGPT, CoMo, and MoMask—on the HumanML3D and KIT benchmarks, the paper reports consistent gains in motion quality, text-motion alignment, and retrieval. If correct, the mechanism offers finer control, such as distinguishing 'left hand' from 'right hand' in a prompt, without changing the underlying generator.

What carries the argument

The load-bearing object is the text-motion aligner: a multi-head attention module that uses each motion token as the query and the full set of text token embeddings as keys and values. It is paired with a composite-aware text encoder that outputs per-token embeddings from a pretrained text encoder instead of a single [CLS] vector. The aligner is what converts a static global condition into a dynamic, per-frame condition, and it is the component that the paper claims carries the improvement.

What would settle it

Train a parameter-matched control in which the same added aligner layers and projections are fed a fixed-length [CLS] condition, for example the same vector repeated at every position, instead of per-token embeddings. If this control matches CASIM's R-Precision and FID gains, the token-level alignment mechanism is not what is doing the work; if it falls short, the composite-aware injection is supported.

Watch

Extended reading notes

Core claim

The central claim is that fixed-length semantic injection is the bottleneck. Where prior text-to-motion models use the CLIP [CLS] token as a global condition, CASIM preserves token-level embeddings and learns a dynamic correspondence between text tokens and motion tokens through multi-head attention—self-attention when the generator is autoregressive, cross-attention when it is a diffusion decoder—so each motion frame can draw on the words that describe it. The paper reports that this injection mechanism, applied on top of existing models with their hyperparameters unchanged, improves R-Precision, FID, and MM-Distance on HumanML3D and KIT-ML, and that attention visualizations show text tokens such as 'wave' attending to early frames and 'sit' to later frames.

Load-bearing premise

The paper attributes the gains to token-level semantic injection, but because each CASIM-augmented model receives extra attention layers and projections that the baseline lacks, the improvements could in principle come from the added model capacity rather than from the alignment mechanism itself.

Editorial extensions

If this is right

  • Gains transfer across generator families: the paper reports improved R-Precision, FID, and MM-Distance for two diffusion models (MDM, MotionDiffuse) and three autoregressive or tokenized models (T2MGPT, CoMo, MoMask) on HumanML3D and KIT-ML.
  • Prompt differences of a single word, such as 'left hand' versus 'right hand,' become distinguishable in the generated motion, giving users finer control over body side, direction, and action order.
  • External semantic augmentation becomes less necessary: CASIM-CoMo without keyword augmentation beats the keyword-augmented CoMo baseline on R-Precision and FID.
  • In long-term generation via DoubleTake, CASIM improves the quality and text alignment of the generated motion clips, with the largest FID gain at a 20-frame handshake (0.953 to 0.463).
  • Because CASIM changes only the semantic injection, it can be layered onto existing generators without retraining their motion tokenizers or denoisers from scratch.

Reading between the lines

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

  • The paper leaves open whether the gains come from the alignment mechanism or from added parameters: a control giving each baseline the same extra attention layers and projections while keeping fixed-length conditioning would settle this.
  • The BERT-versus-CLIP ablation hints that the approach is not tied to CLIP; if so, stronger contextualized text encoders could be swapped in without redesigning the aligner.
  • The MLD result suggests CASIM's benefit depends on motion representations that keep per-frame or per-token structure, so applying it to latent-space generators would require token-wise motion latents rather than a fixed-length vector.
  • The learned word-frame attention could double as an interpretability and editing tool, letting users see which prompt words drive which motion phases and re-weighting attention to edit motion.
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

4 major / 6 minor

Summary. The paper proposes CASIM, a semantic injection module that replaces fixed-length CLIP [CLS] text embeddings with token-level CLIP embeddings and a learned text-motion attention aligner. It instantiates CASIM for autoregressive generators (T2MGPT, CoMo) and diffusion generators (MDM, MotionDiffuse, MoMask), reports quantitative results on HumanML3D and KIT-ML plus a long-horizon DoubleTake study, and includes attention visualizations as an interpretability check.

Significance. If the central claim holds, CASIM is a useful plug-in contribution: it is architecture-agnostic, the authors release code, and the method produces large improvements on several strong baselines, notably MDM and T2MGPT on R-Precision and MM-Dist. The attention visualization in Appendix D is a valuable mechanistic sanity check. However, the central attribution to composite-aware token injection is not yet isolated from model capacity, and the paper's own Appendix C shows a case where the module hurts text-motion alignment, so the contribution is promising but needs additional controlled experiments and more careful claims.

major comments (4)
  1. [Section 4.1; Section 3.2 Eq. (2); Section 3.3 Eqs. (4)-(5)] The central claim that token-level composite-aware injection, rather than added model capacity, drives the reported gains is not yet supported. CASIM adds a linear projection, extends the conditioning sequence from one [CLS] token to N token embeddings, and inserts MHSA/MHCA blocks into each baseline, while Section 4.1 only states that other settings follow the baseline. No baseline is augmented with an equal-parameter conditioning module that still uses only the fixed-length [CLS] embedding. Please add a parameter-matched control, for example a baseline whose text side is expanded to the same number of tokens and given the same attention/MLP machinery but with those tokens derived from [CLS] or from input-independent learned vectors. Without such a control, the improvements in Tables 3 and 4 are confounded with capacity.
  2. [Table 1 vs. Table 3; Table 5] Baseline numbers for the same method and dataset differ across tables without a clear explanation. MDM is reported with Top1 R-Precision 0.455 in Table 1 but 0.471 in Table 3, and T2MGPT is reported with 0.491 in Table 1 but 0.484 in Table 3. Table 5 suggests these correspond to different MDM diffusion-step settings and T2MGPT teacher-forcing settings, but the paper never states which configuration is used in Table 1 versus Table 3, so the reader cannot verify that the CASIM comparison is against the same baseline configuration. Please clarify the configuration used in each table, and report the CASIM-integrated model under the identical baseline setting for every comparison.
  3. [Section 4.2; Tables 3 and 4] The paper claims 'consistent improvements' but reports only point estimates from 20 repeats, with no standard deviations, confidence intervals, or significance tests. Several gains are small enough to be within run-to-run noise, such as MoMask FID 0.064 to 0.057 in Table 3. Please provide variance measures or statistical tests for the main comparisons, at least for the headline metrics (R-Precision Top1, FID, MM-Dist).
  4. [Appendix C, Table 9; Abstract; Section 5] The Abstract and Section 5 state that CASIM 'consistently improves' motion quality and text-motion alignment across state-of-the-art methods, but Appendix C shows CASIM-MLD degrades R-Precision (0.469 to 0.452) and MM-Dist (3.282 to 3.389) on HumanML3D. The explanation that fixed-length motion latents limit fine-grained text-motion correspondence is reasonable, but the claim of consistency must be qualified to the tested model families. This limitation also reinforces the need to characterize when the mechanism helps rather than to present it as universally effective.
minor comments (6)
  1. [Section 2.1] There is a typo in 'directionh adds another layer of difficulty' that should read 'direction adds another layer of difficulty.'
  2. [Section 4.3] The word 'nuaunced' should be 'nuanced.'
  3. [Section 5, Limitations] The text says 'it sill relies on motion blending' and should read 'it still relies on motion blending.'
  4. [Section 3.3 vs. Section 4.2] The symbol τ denotes the diffusion step in Eq. (3) but denotes the teacher-forcing coefficient in Section 4.2; the footnote clarifies this, but the notation clash is confusing within the same paper.
  5. [Table 2 vs. Table 4] CoMo's baseline Top1 R-Precision is 0.422 in Table 2 but 0.399 in Table 4 on KIT-ML; if these correspond to different checkpoint or evaluation settings, this should be stated.
  6. [Section 4.3] The word-cloud analysis in Figure 4 is qualitative; including a small quantitative analysis, such as the fraction of test prompts where the most-attended token is an action verb or modifier, would strengthen the claim that attention focuses on motion-critical words.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: CASIM's claims rest on held-out benchmark measurements with external metrics, not on self-referential definitions or fitted predictions.

full rationale

The paper's derivation chain is empirical rather than definitional. The proposed CASIM mechanism (token-level text embeddings plus a learned text-motion aligner) is integrated into existing generators, and all headline claims are supported by measured R-Precision, FID, MM-Dist, and Diversity scores on held-out test splits of HumanML3D and KIT-ML, computed with the external text-motion matching network of Guo et al. (2022a). There is no equation in which a predicted quantity is defined in terms of the target metric, and no fitted parameter is later renamed as a prediction. The ablation tables (Tables 5, 7, 8) and the self-reported MLD limitation (Appendix C, Table 9) show that CASIM's effect varies with architecture, text encoder, embedding layer, and motion representation, including cases where it slightly hurts text-motion alignment (MLD: R-Precision 0.469 to 0.452). Such variability confirms that the central outcome is not forced by construction. The self-citations in the related-work and acknowledgement sections (Chang et al. 2022, 2023, 2024a, 2024b) describe prior work on gesture synthesis, affect conditioning, and synthetic group activities; none of these citations supplies a load-bearing premise for CASIM's claimed improvement. The reviewer concern that baselines are not given matched parameter counts is a legitimate external-validity and capacity-confounding question, but it is a correctness risk, not a circularity. The paper's central claim is therefore not equivalent to its inputs: it is a testable empirical result whose sign and magnitude are contingent on the underlying model, representation, and dataset.

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

The central claim rests on the assumption that token-level CLIP embeddings retain the composite semantics the authors want to inject, and that the attention-based aligner is the cause of the observed gains. No additional entities or fitted constants are introduced. The main unverified premise is that the improvements are attributable to the injection mechanism rather than to the added parameters of the aligner.

assumptions (4)
  • domain assumption Pretrained CLIP text encoder token embeddings preserve composable semantic details needed for motion alignment.
    Section 3.1 assumes token-level embeddings from CLIP retain the granular distinctions that the [CLS] vector loses.
  • domain assumption The text-to-motion evaluation metrics (R-Precision, FID, MM-Dist) computed with the Guo et al. matching network reliably measure generation quality and alignment.
    Section 4.1 relies on this external evaluation protocol without auditing it.
  • domain assumption Adding the text-motion aligner to each baseline does not negatively interact with the base model's training objective or architecture.
    Section 4.1 applies CASIM by replacing fixed-length text injection with token-level attention, assuming the original training schedules remain valid.
  • domain assumption Autoregressive and diffusion generators can benefit from per-token cross-attention conditioning without architectural redesign.
    Sections 3.2 and 3.3 assume the attention-based aligner integrates cleanly into both generation paradigms.

how reviews work

0 comments
Cite this review

Pith. "Pith review of CASIM: Composite Aware Semantic Injection for Text to Motion Generation." pith.science (2026). https://pith.science/paper/PORDPCX5

@misc{pith2026250202063,
  author       = {Pith},
  title        = {Pith review of: CASIM: Composite Aware Semantic Injection for Text to Motion Generation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/PORDPCX5}},
  note         = {Machine review of arXiv:2502.02063}
}
read the original abstract

Recent advances in generative modeling and tokenization have driven significant progress in text-to-motion generation, leading to enhanced quality and realism in generated motions. However, effectively leveraging textual information for conditional motion generation remains an open challenge. We observe that current approaches, primarily relying on fixed-length text embeddings (e.g., CLIP) for global semantic injection, struggle to capture the composite nature of human motion, resulting in suboptimal motion quality and controllability. To address this limitation, we propose the Composite Aware Semantic Injection Mechanism (CASIM), comprising a composite-aware semantic encoder and a text-motion aligner that learns the dynamic correspondence between text and motion tokens. Notably, CASIM is model and representation-agnostic, readily integrating with both autoregressive and diffusion-based methods. Experiments on HumanML3D and KIT benchmarks demonstrate that CASIM consistently improves motion quality, text-motion alignment, and retrieval scores across state-of-the-art methods. Qualitative analyses further highlight the superiority of our composite-aware approach over fixed-length semantic injection, enabling precise motion control from text prompts and stronger generalization to unseen text inputs.

Figures

Figures reproduced from arXiv: 2502.02063 by the authors.

Figure 1
Figure 1. (Top) Fixed-length semantic injection, which primarily relied on the [CLS] token embedding from CLIP (Radford et al., 2021) to represent the entire text prompt, fails to capture the subtle differences in individ￾ual words. As a result, it generates highly similar motions from distinct text prompts. (Bottom) Our Composite aware semantic injection method allows each motion frame to dy￾namically attend to every word to… view at source ↗
Figure 2
Figure 2. CASIM consists of two major components: Composite Aware Text Encoder (Left) for extracting granular [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Qualitative comparison between two baselines, their CASIM-enhanced models, and ground truth (GT) on [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figures from the paper (2 more)
Figure 4
Figure 4. Figure 4: Analysis of attention patterns in CASIM. Left: [PITH_FULL_IMAGE:figures/full_fig_p007_4.png]
Figure 5
Figure 5. Figure 5: Visualization of attention weights in CASIM-MDM. Top: Generated motion sequence for the prompt ” [PITH_FULL_IMAGE:figures/full_fig_p012_5.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

40 extracted references · 26 canonical work pages

  1. [1]

    write newline

    " write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 global.max substring 't := if while FUNCTION format.date year duplicate empty "emp...

  2. [2]

    Alexanderson, S., Nagy, R., Beskow, J., and Henter, G. E. Listen, denoise, action! audio-driven motion synthesis with diffusion models. ACM Trans. Graph., 42 0 (4): 0 44:1--44:20, 2023. doi:10.1145/3592458

    doi:10.1145/3592458 2023

  3. [3]

    J., and Varol, G

    Athanasiou, N., Cseke, A., Diomataris, M., Black, M. J., and Varol, G. Motionfix: Text-driven 3d human motion editing. In SIGGRAPH Asia 2024 Conference Proceedings. ACM, December 2024. URL https://motionfix.is.tue.mpg.de/

    work page 2024

  4. [4]

    Generating human motion in 3d scenes from text descriptions

    Cen, Z., Pi, H., Peng, S., Shen, Z., Yang, M., Shuai, Z., Bao, H., and Zhou, X. Generating human motion in 3d scenes from text descriptions. In CVPR, 2024

    work page 2024

  5. [5]

    The ivi lab entry to the genea challenge 2022--a tacotron2 based method for co-speech gesture generation with locality-constraint attention mechanism

    Chang, C.-J., Zhang, S., and Kapadia, M. The ivi lab entry to the genea challenge 2022--a tacotron2 based method for co-speech gesture generation with locality-constraint attention mechanism. In Proceedings of the 2022 International Conference on Multimodal Interaction, pp.\ 784--789, 2022

    work page 2022

  6. [6]

    S., Zhang, S., Jayashankar, R., Usman, M., and Kapadia, M

    Chang, C.-J., Sohn, S. S., Zhang, S., Jayashankar, R., Usman, M., and Kapadia, M. The importance of multimodal emotion conditioning and affect consistency for embodied conversational agents. In Proceedings of the 28th International Conference on Intelligent User Interfaces, pp.\ 790--801, 2023

    work page 2023

  7. [7]

    On the equivalency, substitutability, and flexibility of synthetic data

    Chang, C.-J., Li, D., Moon, S., and Kapadia, M. On the equivalency, substitutability, and flexibility of synthetic data. arXiv preprint arXiv:2403.16244, 2024 a

    arXiv 2024

  8. [8]

    S., Yoon, S., Pavlovic, V., and Kapadia, M

    Chang, C.-J., Li, D., Patel, D., Goel, P., Zhou, H., Moon, S., Sohn, S. S., Yoon, S., Pavlovic, V., and Kapadia, M. Learning from synthetic human group activities. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp.\ 21922--21932, 2024 b

    work page 2024

Show all 40 references
  1. [9]

    Motionllm: Understanding human behaviors from human motions and videos

    Chen, L.-H., Lu, S., Zeng, A., Zhang, H., Wang, B., Zhang, R., and Zhang, L. Motionllm: Understanding human behaviors from human motions and videos. arXiv preprint arXiv:2405.20340, 2024

    2024 arXiv

  2. [10]

    Executing your commands via motion diffusion in latent space

    Chen, X., Jiang, B., Liu, W., Huang, Z., Fu, B., Chen, T., and 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

    2023

  3. [11]

    PoseScript: 3D Human Poses from Natural Language

    Delmas, G., Weinzaepfel, P., Lucas, T., Moreno-Noguer, F., and Rogez, G. PoseScript: 3D Human Poses from Natural Language . In ECCV , 2022

    2022

  4. [12]

    Bert: Pre-training of deep bidirectional transformers for language understanding

    Devlin, J. Bert: Pre-training of deep bidirectional transformers for language understanding. arXiv preprint arXiv:1810.04805, 2018

    2018 arXiv

  5. [13]

    and Nichol, A

    Dhariwal, P. and Nichol, A. Diffusion models beat gans on image synthesis. Advances in neural information processing systems, 34: 0 8780--8794, 2021

    2021

  6. [14]

    K., and Fatahalian, K

    Goel, P., Wang, K.-C., Liu, C. K., and Fatahalian, K. Iterative motion editing with natural language. In ACM SIGGRAPH 2024 Conference Papers, pp.\ 1--9, 2024

    2024

  7. [15]

    Action2motion: Conditioned generation of 3d human motions

    Guo, C., Zuo, X., Wang, S., Zou, S., Sun, Q., Deng, A., Gong, M., and Cheng, L. Action2motion: Conditioned generation of 3d human motions. In Proceedings of the 28th ACM International Conference on Multimedia, pp.\ 2021--2029, 2020

    2021

  8. [16]

    Generating diverse and natural 3d human motions from text

    Guo, C., Zou, S., Zuo, X., Wang, S., Ji, W., Li, X., and Cheng, L. Generating diverse and natural 3d human motions from text. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pp.\ 5152--5161, June 2022 a

    2022

  9. [17]

    Tm2t: Stochastic and tokenized modeling for the reciprocal generation of 3d human motions and texts

    Guo, C., Zuo, X., Wang, S., and Cheng, L. Tm2t: Stochastic and tokenized modeling for the reciprocal generation of 3d human motions and texts. In European Conference on Computer Vision, pp.\ 580--597. Springer, 2022 b

    2022

  10. [18]

    G., Wang, S., and Cheng, L

    Guo, C., Mu, Y., Javed, M. G., Wang, S., and Cheng, L. Momask: Generative masked modeling of 3d human motions. 2023

    2023

  11. [19]

    Como: Controllable motion generation through language guided pose code editing, 2024

    Huang, Y., Wan, W., Yang, Y., Callison-Burch, C., Yatskar, M., and Liu, L. Como: Controllable motion generation through language guided pose code editing, 2024

    2024

  12. [20]

    Motiongpt: Human motion as a foreign language

    Jiang, B., Chen, X., Liu, W., Yu, J., Yu, G., and Chen, T. Motiongpt: Human motion as a foreign language. Advances in Neural Information Processing Systems, 36, 2024

    2024

  13. [21]

    Act as you wish: Fine-grained control of motion diffusion model with hierarchical semantic graphs

    Jin, P., Wu, Y., Fan, Y., Sun, Z., Wei, Y., and Yuan, L. Act as you wish: Fine-grained control of motion diffusion model with hierarchical semantic graphs. In NeurIPS, 2023

    2023

  14. [22]

    Guided motion diffusion for controllable human motion synthesis

    Karunratanakul, K., Preechakul, K., Suwajanakorn, S., and Tang, S. Guided motion diffusion for controllable human motion synthesis. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pp.\ 2151--2162, 2023

    2023

  15. [23]

    M., and Rogez, G

    Lee, T., Baradel, F., Lucas, T., Lee, K. M., and Rogez, G. T2lm: Long-term 3d human motion generation from multiple sentences. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp.\ 1867--1876, 2024

    2024

  16. [24]

    Unimotion: Unifying 3d human motion synthesis and understanding

    Li, C., Chibane, J., He, Y., Pearl, N., Geiger, A., and Pons-Moll, G. Unimotion: Unifying 3d human motion synthesis and understanding. arXiv preprint arXiv:2409.15904, 2024

    2024 arXiv

  17. [25]

    F., Pons-Moll, G., and Black, M

    Mahmood, N., Ghorbani, N., Troje, N. F., Pons-Moll, G., and 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

    2019

  18. [26]

    Gpt-4 technical report, 2024

    OpenAI, Achiam, J., Adler, S., et al. Gpt-4 technical report, 2024. URL https://arxiv.org/abs/2303.08774

    2024 arXiv

  19. [27]

    The KIT motion-language dataset

    Plappert, M., Mandery, C., and Asfour, T. The KIT motion-language dataset. Big Data, 4 0 (4): 0 236--252, dec 2016. doi:10.1089/big.2016.0028. URL http://dx.doi.org/10.1089/big.2016.0028

    2016

  20. [28]

    Improving language understanding by generative pre-training

    Radford, A., Narasimhan, K., Salimans, T., and Sutskever, I. Improving language understanding by generative pre-training. 2018

    2018

  21. [29]

    W., Hallacy, C., Ramesh, A., Goh, G., Agarwal, S., Sastry, G., Askell, A., Mishkin, P., Clark, J., et al

    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

    2021

  22. [30]

    B., and van de Panne, M

    Setareh, C., Tevet, G., Reda, D., Peng, X. B., and van de Panne, M. Generating human interaction motions in scenes with text control. 2024

    2024

  23. [31]

    Shafir, Y., Tevet, G., Kapon, R., and Bermano, A. H. Human motion diffusion as a generative prior. In The Twelfth International Conference on Learning Representations, 2024

    2024

  24. [32]

    Generating fine-grained human motions using chatgpt-refined descriptions

    Shi, X., Luo, C., Peng, J., Zhang, H., and Sun, Y. Generating fine-grained human motions using chatgpt-refined descriptions. arXiv preprint arXiv:2312.02772, 2023

    2023 arXiv

  25. [33]

    Score-guided diffusion for 3d human recovery

    Stathopoulos, A., Han, L., and Metaxas, D. Score-guided diffusion for 3d human recovery. In CVPR, 2024

    2024

  26. [34]

    Tevet, G., Raab, S., Gordon, B., Shafir, Y., Cohen-or, D., and Bermano, A. H. Human motion diffusion model. In The Eleventh International Conference on Learning Representations, 2023. URL https://openreview.net/forum?id=SJ1kSyO2jwu

    2023

  27. [35]

    Neural discrete representation learning

    Van Den Oord, A., Vinyals, O., et al. Neural discrete representation learning. Advances in neural information processing systems, 30, 2017

    2017

  28. [36]

    T2m-gpt: Generating human motion from textual descriptions with discrete representations

    Zhang, J., Zhang, Y., Cun, X., Huang, S., Zhang, Y., Zhao, H., Lu, H., and Shen, X. T2m-gpt: Generating human motion from textual descriptions with discrete representations. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2023 a

    2023

  29. [37]

    Motiondiffuse: Text-driven human motion generation with diffusion model

    Zhang, M., Cai, Z., Pan, L., Hong, F., Guo, X., Yang, L., and Liu, Z. Motiondiffuse: Text-driven human motion generation with diffusion model. arXiv preprint arXiv:2208.15001, 2022

    2022 arXiv

  30. [38]

    Finemogen: Fine-grained spatio-temporal motion generation and editing

    Zhang, M., Li, H., Cai, Z., Ren, J., Yang, L., and Liu, Z. Finemogen: Fine-grained spatio-temporal motion generation and editing. NeurIPS, 2023 b

    2023

  31. [39]

    Zhao, L., Peng, X., Tian, Y., Kapadia, M., and Metaxas, D. N. Semantic graph convolutional networks for 3d human pose regression. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pp.\ 3425--3435, 2019

    2019

  32. [40]

    Avatargpt: All-in-one framework for motion understanding planning generation and beyond

    Zhou, Z., Wan, Y., and Wang, B. Avatargpt: All-in-one framework for motion understanding planning generation and beyond. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp.\ 1357--1366, 2024

    2024

Pith tools

Reviewed August 9, 2026 · model on record in the stance chip above.