Pith. sign in

REVIEW 4 major objections 6 minor 196 references

CustomDance: Customized 3D Dance Generation with Coarse-to-Fine Human-Centered Interactive Control

T0 review · 4 major / 6 minor · reviewed 2026-08-10 · deepseek-v4-flash

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

Pith's one-line read CustomDance claims that 3D dance generation is better done as a three-stage, human-in-the-loop authoring workflow—MLLM anchor planning, multimodal phrase retrieval, and diffusion-based gap filling and repair—than as a single end-to-end…

desk verdict A well-built interactive choreography system with a genuine three-stage workflow; the headline quantitative claim is partly confounded by library retrieval, but the user-study and long-sequence results carry the paper. read the letter →

arxiv 2608.06722 v1 pith:EY2D24WV submitted 2026-08-07 cs.HC

Xulong Tang , Kaixing Yang , Xiaohu Guo , Prabhakaran Balakrishnan , Rawan Alghofaili This is my paper · ORCID
classification cs.HC
keywords 3Ddancegenerationhuman-in-the-loopauthoringmultimodalretrievaldiffusioninpaintingmusic-conditionedmotionsynthesislargelanguagemodelchoreographyqualityevaluation
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

CustomDance is an interactive system for authoring 3D dances, built around the claim that choreography is best treated as a human-in-the-loop process rather than a one-shot music-to-motion mapping. The paper proposes a three-stage workflow that mirrors how choreographers actually work: a multimodal language model reads the music and a global text prompt to propose where phrases should go and what should happen there; a retriever ranks short motion clips by local music, text, and body-part controls; and a diffusion model sews the chosen clips together and repairs flagged glitches. The system's evidence is that non-expert users can author a 32-second dance in about twelve minutes with no leftover kinematic artifacts, and that blinded raters prefer the results over two recent generative baselines in the large majority of comparisons. If the claim holds, the practical effect is that AI assistance for dance moves from generating something plausible to helping the user make what they actually want, with the human deciding structure and style at each step.

What carries the argument

The argument runs on four engineered objects working together. The anchor planner is an MLLM prompted to output JSON anchor–cue pairs that respect musical structure and a minimum temporal separation, converting abstract intent into concrete 4-second slots. The retriever is the search instrument: a contrastive trimodal model in which concatenated text and UI-control embeddings gate the music embedding through FiLM layers, trained with InfoNCE loss and modality dropout at 0.2 so missing inputs degrade gracefully, with library embeddings precomputed offline so each query costs about 0.7 seconds. The generative workhorse is a DDIM diffusion model trained on 8-second clips with an x0-prediction loss plus kinematic auxiliary losses, and its masked-denosing update—keeping known frames fixed while resynthesizing masked ones—implements both gap completion and joint-level repair without additional training. Finally, the diagnoser computes lightweight kinetic-energy proxies, absolute kinetic energy from linear joint velocities and relative kinetic energy from SO(3) angular velocities via the logarithm map, across six joint groups, so users can see exactly where root teleportation or limb twists occur and repair only those spans.

What would settle it

Drop the roughly 3 hours of PromptHMR-reconstructed video and the StableMotion augmentations from the library, retrain the retriever and diffusion model, and re-run the Study 2 metrics on the same five 32-second clips; if FID_k and Div_k stay near 22.55 and 7.02 the reconstructed portion is not load-bearing, and if they jump toward MEGADance's 32.59 the headline quality numbers rest on data the paper never verifies against ground-truth motion. A second check: re-run the retriever on held-out phrases with the text modality masked off and compare Recall@10 to the reported 72.97, which isolates whether the claimed multimodal conditioning actually contributes.

Watch

Extended reading notes

Core claim

The paper's central claim is that high-quality customized dance generation is an authoring problem, not just a synthesis problem, and that the right unit of design is a three-stage, coarse-to-fine pipeline. In the first stage, an MLLM (Gemini) analyzes the global music and a high-level text description and returns temporal anchors with imperative creative cues, which become 4-second phrase slots on a timeline. In the second, a contrastively trained trimodal retriever—jointly embedding local music (Librosa features), local text (CLIP), and six joint-group intensity/variety controls—ranks candidate phrases from a roughly 10.7-hour curated library (7.7 hours of FineDance motion capture, about 3 hours of PromptHMR-reconstructed video, augmented with StableMotion) and returns a top-10 list the user can preview and accept. In the third, a DDIM music-conditioned diffusion inpainter, built on a BiMamba–Transformer backbone, fills unassigned gaps by masked denoising and repairs user-selected intervals flagged by kinetic-energy visualizations across six joint groups, iterating until the user is satisfied. The paper reports that this workflow cuts average authoring time from 28.41 to 11.66 minutes and clip replacements from 25.16 to 5.32 versus a timeline-only editor, eliminates detector-flagged artifacts, and achieves FID 22.55, diversity 7.02/6.44, and BAS 0.233, beating Lodge, MEGADance, and the baseline editor on all objective metrics and earning 82–93% rank-1 preference rates in a blinded comparison.

Load-bearing premise

The whole system—the retrieved phrases and the diffusion model alike—rests on the curated 10.7-hour dance library being clean, consistently aligned, and representative of the styles users request; if noise from the reconstructed video, coordinate mismatches, or augmentation artifacts seep in, both the authoring experience and the objective metrics change regardless of how good the interaction design is.

Editorial extensions

If this is right

  • Non-dancers can author a complete 32-second choreography in about 12 minutes with zero detector-flagged kinematic artifacts, down from 28+ minutes and more than five artifacts with a plain timeline editor.
  • For 90-second pieces, active authoring with CustomDance takes about 31 minutes, roughly half the time of professional scratch choreography, while expert raters judge the result close to the professionals on pose satisfaction and description fulfillment.
  • Blinded raters rank CustomDance dances first 82–93% of the time across music alignment, description fulfillment, and overall performance when compared against Lodge and MEGADance, so the authoring paradigm, not just the generator, is what users prefer.
  • The objective metrics (FID 22.55, diversity 7.02/6.44, BAS 0.233) suggest the composed dances are more realistic, more varied, and better beat-aligned than end-to-end baselines, which would make retrieval-plus-inpainting a competitive recipe for music-conditioned motion.

Reading between the lines

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

  • My inference: the biggest measured win is search efficiency, not synthesis quality—authoring time dropped 59% when ranking replaced random browsing (28.28 to 11.66 minutes), so the retriever's ranking likely carries more of the user-facing value than the diffusion repair; a direct test would replace the learned retriever with music-similarity-only ranking and measure the drop.
  • The fixed 4-second slot follows eight-count phrasing at roughly 120 BPM, and nothing in the system adapts slot length to the music's tempo; a testable extension would make the slot duration tempo-aware and measure whether phrase appropriateness rises for slow or fast tracks.
  • Because final dances are mostly assembled from retrieved clips, the reported diversity partly measures the library's phrase coverage, which suggests the cheapest path to more diverse output is adding style-balanced motion capture rather than changing the generator.
  • The plan-retrieve-inpaint decomposition is not dance-specific; it could be tested as a general authoring paradigm for any timeline-based embodied content, such as gesture for virtual agents, martial-arts sequences, or character animation, where the user holds local creative intent.
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. CustomDance presents a three-stage interactive choreography system: an MLLM (Gemini) generates temporal anchors and creative cues from music and a user text prompt; a trimodal retriever ranks 4-second phrases from a 10.7-hour dance library conditioned on local music, text, and joint-group intensity/variety controls; and a diffusion inpainter fills gaps and repairs localized artifacts. The paper reports a 25-participant within-subject authoring study (Study 1), a blinded 25-rater quality ranking against Lodge, MEGADance, and a baseline editor (Study 2), and a 90-second long-sequence study with expert raters and professional scratch choreography (Study 3), along with component-level evaluations. The central claim is that the full system improves authoring efficiency, perceived control, and final dance quality over the ablated conditions and over end-to-end generation baselines.

Significance. If the results hold, the paper is a meaningful contribution to AI-assisted choreography and interactive character animation: it provides a concrete instantiation of the choreographer-inspired workflow, with a clear interface design and a thorough user-study protocol. Strengths include the within-subject ablation design, the blinded ranking procedure, the long-sequence comparison that includes professional scratch choreography, and the honest Limitations section that already concedes the retrieval-based metric concern. The authoring-time reductions (11.66 vs 28.41 minutes) and high rank-1 rates are impressive and support the central authoring-benefit claim. However, the objective metric comparison and the artifact-count measure have confounds that must be resolved before the quantitative superiority claims can be accepted as stated. The circularity burden of the paper is low because it makes no formal derivation, and the free parameters (slot duration, candidate list size, CFG scale, dropout probabilities) are disclosed explicitly.

major comments (4)
  1. [§7.3, Table 2] The objective comparison is confounded by retrieval and human selection. CustomDance outputs are composed largely of user-selected motion-capture phrases from a curated library, while Lodge and MEGADance synthesize from learned priors. FID, Diversity, and BAS therefore measure different objects: retrieval-aided composition versus full generation. The paper's own Limitations section (§8) concedes that 'fidelity-oriented metrics may partially favor retrieval-supported outputs.' Since the abstract and Section 7.3.2 explicitly claim superiority based on these numbers, Table 2 does not establish that the proposed generative and planning components improve motion fidelity. I recommend either reporting metrics on the diffusion-inpainted (gap and repair) portions only, or giving Lodge and MEGADance access to the same retrieval library, or re-labeling the comparison as a system-level study rather than a method-level one.
  2. [§7.2.1–§7.2.2, Table 1] The artifact-count metric appears circular with the Diagnoser. The artifact count is defined as 'thresholded kinematic anomalies,' and the Diagnoser uses the same class of AKE/RKE kinematic proxies. Unsurprisingly, conditions with the Diagnoser enabled report 0.00 artifacts. The footnote that the offline metric is 'separate from the AKE/RKE curves' is not sufficient, because both target the same kinematic outliers (root jumps, limb twists). Please provide an independent artifact definition (e.g., foot skating, penetration, joint-angle limit violations) or demonstrate that the offline metric was not guiding the repair actions in a way that trivially removes exactly those anomalies.
  3. [§7.3, Table 2] No error bars, confidence intervals, or significance tests are reported for any objective metric. It is unclear how many generations or seeds were used and whether the FID, Diversity, and BAS values are statistically distinguishable. Please report mean and standard deviation over at least five independent runs (or over the five selected dances) and apply a statistical test when comparing methods.
  4. [§7.1] The dance library combines 7.7 hours of FineDance motion capture, approximately 3 hours of PromptHMR-reconstructed video, and StableMotion augmentations, with no ground-truth validation of the reconstructed or augmented motions. If HMR reconstruction noise, coordinate misalignment after the FineDance alignment, or augmentation artifacts enter the library, both the retrieval quality and the objective metrics in Table 2 would be affected independently of the interaction design. The paper should report validation (e.g., reconstruction error on a held-out set, or manual inspection rates) or an ablation showing that retrieval quality and objective metrics are stable to the HMR/augmented subset.
minor comments (6)
  1. [§6.3, Eq. (5)] Equation (5) introduces the masked denoising update, but the notation q(x_known, t-1) is not defined; please clarify the forward-process conditional and the noise schedule used.
  2. [§7.4, Table 3] Table 3 reports FSR and Jitter without in-text definitions; a brief definition or a precise pointer to the original sources would help the reader interpret the long-sequence results.
  3. [§7.2.1] The artifact-count metric is only described as 'thresholded kinematic anomalies' with details deferred to the supplementary material; a concise definition in the main text is needed to assess the circularity concern raised in the major comments.
  4. [§7.3.1] The 'same repair workflow' applied to Baseline, Lodge, and MEGADance is not specified; please describe the repair protocol, including who invoked it and what constraints were observed, so that the fairness of the comparison can be evaluated.
  5. [§4] The behavior of overlapping slots ('Selecting one temporarily invalidates conflicting slots') is confusing; please clarify whether the conflicting slot's content is preserved and how the user's undo action restores it.
  6. [References] Some references are not yet published (e.g., MEGADance [NeurIPS'25], OmniDance [2026b]); if final versions exist, they should be cited to help readers verify the baselines.

Circularity Check

2 steps flagged · score 4.0 of 10

Objective-metric superiority is partly confounded by retrieval from the reference-motion distribution, and the artifact-count ablation is close to measuring the same kinematic anomalies the Diagnoser repairs; the core interactive-authoring claims remain independent.

  1. other [§7.1 Dance Library/Training Clips, §7.3.2 Objective Results, §8 Limitations]
    "CustomDance obtains the lowest FID_k (22.55) and the highest Div_k (7.02), Div_g (6.44), and BAS (0.233) among the compared methods, indicating strong motion fidelity... Limitations: ... fidelity-oriented metrics may partially favor retrieval-supported outputs."

    CustomDance fills phrase slots by retrieving real 4-second motion clips from a curated library of FineDance mocap and HMR-reconstructed video, while Lodge and MEGADance synthesize from noise or learned priors. FID, Diversity, and BAS are computed against the same class of real-motion reference data, so a retrieval system that returns near-reference samples will score lower FID and higher BAS by construction, independent of whether the MLLM planning, retriever, or diffusion inpainting improves synthesis. The paper's own limitation concedes the confound. This makes the Table 2 objective advantage load-bearing only for the retrieval-plus-library system, not for the generative components; the Study 1 authoring-time and perception results are not affected by this issue.

  2. other [§7.2.1 Ablation Study Design and §7.2.2 Authoring Experience Results, cf. §6.2 Artifact Detection and Repair]
    "The artifact count is an offline objective counting metric based on thresholded kinematic anomalies, separate from the AKE/RKE curves used for interactive Diagnoser visualization... conditions with the Diagnoser enabled show zero detector-flagged artifacts related to local kinematic discontinuities."

    The Diagnoser visualizes AKE/RKE kinetic-energy spikes to help users locate 'abnormal intervals such as single-frame teleportation or abnormal limb twists,' and the Remaker repairs exactly those flagged regions. The offline artifact-count metric is also 'based on thresholded kinematic anomalies,' so the claim that Diagnoser-guided refinement reduces artifact counts is partly measuring the same anomaly class that was repaired. The word 'separate' softens the reduction, but the scoring target and the repair target remain the same kinematic-discontinuity concept, making the artifact-count comparison closer to a reconstruction of the repair objective than an independent quality evaluation.

full rationale

The paper is an empirical systems paper rather than a derivation chain, and no load-bearing argument reduces to a self-citation or to a fitted parameter renamed as a prediction. The self-authored citations (e.g., MEGADance for the BiMamba-Transformer architecture, Tang et al. for intensity/variety controls) are implementation influences, not forced conclusions. The two concerns above are real but partial: the objective-quality advantage in Table 2 is confounded because CustomDance's outputs are largely retrieved real-motion phrases drawn from the same distribution used by FID/Diversity/BAS, a point the paper itself acknowledges in §8; and the artifact-count ablation in Table 1 uses a thresholded kinematic-anomaly metric that overlaps with the anomaly class the Diagnoser is designed to expose. Neither issue undermines the central interactive-authoring claim, which rests on within-system ablations, user-rating comparisons, and timing measurements that are independent of the retrieval-reference confound. The overall circularity is therefore moderate but not severe: the system's core contribution is not forced by its inputs, but part of the quantitative superiority claim is substantially self-supplied by the retrieval library.

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

The contribution is an interactive system rather than a mathematical derivation, so the ledger is dominated by empirical and design assumptions: the quality and alignment of the dance library, the reliability of the closed planning API, and the validity of the evaluation metrics. The free parameters are design choices, several validated by component studies, but none are fitted constants in a derivation. No new physical or conceptual entities are introduced.

free parameters (6)
  • Slot duration Delta_slot = 4 seconds
    Retrieval phrase length set by design; a component study supports 4s over 2s and 8s, but this choice determines what the retriever and inpainter receive.
  • Candidate list size K = 10
    Top-10 candidates chosen after a preview-usability study; this affects user effort and the quality of the final selection.
  • Classifier-free guidance scale = 2.5
    Inference guidance scale for the diffusion generator, chosen by standard practice and not ablated in the main text.
  • Modality dropout probability = 0.2
    Retriever training dropout for text and UI modalities; chosen for robustness but not ablated in the main text.
  • Music conditioning dropout = 0.25
    Diffusion training dropout for classifier-free guidance; standard value and not ablated in the main text.
  • Default intensity and variety controls = 0.5
    UI control defaults used when the user does not adjust the sliders; these values directly influence retrieved phrases.
assumptions (6)
  • domain assumption FineDance motion capture data and its genre annotations are accurate enough to serve as the retrieval library and training corpus.
    The quality of the library is the substrate for all retrieval and inpainting results, and no independent verification of the 16-genre annotations is provided.
  • domain assumption PromptHMR human mesh recovery from internet dance videos yields SMPL motion clean enough for retrieval after coordinate alignment.
    The 3 hours of HMR-reconstructed motion are curated but not validated against ground truth, so reconstruction noise would flow into the library.
  • domain assumption StableMotion augmentation preserves choreographic quality and style while expanding the corpus.
    Augmented clips are treated as equivalent to real motion for training and retrieval, but no evaluation of augmentation fidelity is reported.
  • domain assumption The Gemini MLLM with the structured prompt reliably extracts musical anchors and creative cues from audio and text.
    The system depends on a closed API, and its reliability is asserted by an expert rating study but the API and prompt are not reproducible.
  • standard math Masked denoising under DDIM can inpaint missing temporal gaps and repaired joints while preserving known motion.
    Equation (5) is the standard diffusion inpainting operator, and the paper assumes it produces seamless dances without retraining.
  • domain assumption The objective metrics and the thresholded kinematic artifact count capture choreographic quality as experienced by users.
    Metrics such as FID, diversity, BAS, FSR, and Jitter are accepted in the music-to-dance literature but are proxies, and artifact thresholds are not fully specified in the main text.

how reviews work

0 comments
Cite this review

Pith. "Pith review of CustomDance: Customized 3D Dance Generation with Coarse-to-Fine Human-Centered Interactive Control." pith.science (2026). https://pith.science/paper/EY2D24WV

@misc{pith2026260806722,
  author       = {Pith},
  title        = {Pith review of: CustomDance: Customized 3D Dance Generation with Coarse-to-Fine Human-Centered Interactive Control},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/EY2D24WV}},
  note         = {Machine review of arXiv:2608.06722}
}
read the original abstract

With the rise of AI-generated content (AIGC) and advanced techniques for 3D human representation, the task of generating 3D dance movements has become an exciting area of research. Despite significant advancements, current methods often fail to provide comprehensive and distinct control over various multimodal inputs from users, such as music or specific descriptions of desired movements. As a result, the generated motions may be statistically plausible and technically correct, but they often lack depth, expressiveness, and alignment with the user's creative vision. To address this issue, we present CustomDance, a coarse-to-fine interactive system designed for customized 3D dance generation. Inspired by the workflows of expert choreographers, CustomDance introduces a novel paradigm to AI-assisted choreography through three interconnected stages. First, a multimodal Large Language Model (MLLM) analyzes the music and a high-level text prompt to identify key temporal anchors and creative cues for the piece. Next, for each anchor, a multimodal retriever suggests high-quality motion clips from a dance library based on local music and text, empowering the user with concrete and predictable options. Finally, a custom music-conditioned diffusion in-painter seamlessly connects the selected phrases, allowing for iterative, user-guided refinement of the final composition, supported by visualizations of motion dynamics. Our evaluations demonstrate that CustomDance not only highlights the significant creative utility and empowering potential of our AI-assisted choreography paradigm, but also outperforms competitive baselines across quantitative and qualitative comparisons.

Figures

Figures reproduced from arXiv: 2608.06722 by the authors.

Figure 1
Figure 1. Users provide music and choreography preferences. Prior dance generation methods often fail to satisfy these preferences and frequently exhibit [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Overview of CustomDance . 2.2 Interactive 3D Motion Authoring To enable creators to design 3D movements without directly manip￾ulating low-level joint trajectories, interactive 3D motion authoring systems and control operators have been widely studied in computer graphics and human-computer interaction [Gou et al. 2025]. Broadly, prior systems can be grouped into four categories: performance￾driven control, example-… view at source ↗
Figure 3
Figure 3. System overview of the interactive choreography pipeline. [PITH_FULL_IMAGE:figures/full_fig_p011_3.png] view at source ↗
Figures from the paper (4 more)
Figure 4
Figure 4. Figure 4: Task-level 7-point Likert ratings on Perceived Controllability, Pose Satisfaction, and Description Fulfillment across the five interface conditions. [PITH_FULL_IMAGE:figures/full_fig_p011_4.png]
Figure 5
Figure 5. Figure 5: Rank distributions of CustomDance, MEGADance, Lodge, and Baseline across Music Alignment, Description Fulfillment, and Overall Performance. [PITH_FULL_IMAGE:figures/full_fig_p011_5.png]
Figure 6
Figure 6. Figure 6: Qualitative results of CustomDance across five FineDance coarse styles (Classic, Folk, Standard, Mix, Street), conditioned on user preferences. [PITH_FULL_IMAGE:figures/full_fig_p012_6.png]
Figure 7
Figure 7. Figure 7: Qualitative comparison on a street-style excerpt conditioned on user preferences. [PITH_FULL_IMAGE:figures/full_fig_p012_7.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

196 extracted references · 24 canonical work pages

  1. [1]

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

    Bailando: 3d dance generation by actor-critic gpt with choreographic memory , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

  2. [2]

    IEEE Transactions on Pattern Analysis and Machine Intelligence , year=

    Bailando++: 3d dance gpt with choreographic memory , author=. IEEE Transactions on Pattern Analysis and Machine Intelligence , year=

  3. [3]

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

    A brand new dance partner: Music-conditioned pluralistic dancing controlled by multiple dance genres , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

  4. [4]

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

    Edge: Editable dance generation from music , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

  5. [5]

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

    Lodge: A coarse to fine diffusion network for long dance generation guided by the characteristic dance primitives , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

  6. [6]

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

    Optimizing Diffusion Noise Can Serve as Universal Motion Priors , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

  7. [7]

    Proceedings of the 2024 International Conference on Multimedia Retrieval , pages=

    CoDancers: Music-Driven Coherent Group Dance Generation with Choreographic Unit , author=. Proceedings of the 2024 International Conference on Multimedia Retrieval , pages=

    2024

  8. [8]

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

    Music-driven group choreography , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

Show all 196 references
  1. [9]

    ACM Transactions on Graphics (TOG) , volume=

    Controllable group choreography using contrastive diffusion , author=. ACM Transactions on Graphics (TOG) , volume=. 2023 , publisher=

    2023

  2. [10]

    arXiv preprint arXiv:2403.18811 , year=

    Duolando: Follower gpt with off-policy reinforcement learning for dance accompaniment , author=. arXiv preprint arXiv:2403.18811 , year=

    arXiv

  3. [11]

    Proceedings of the IEEE/CVF International Conference on Computer Vision , pages=

    Ai choreographer: Music conditioned 3d dance generation with aist++ , author=. Proceedings of the IEEE/CVF International Conference on Computer Vision , pages=

  4. [12]

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

    Score-Guided Diffusion for 3D Human Recovery , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

  5. [13]

    Proceedings of the IEEE/CVF International Conference on Computer Vision , pages=

    Tm2d: Bimodality driven 3d dance generation via music-text integration , author=. Proceedings of the IEEE/CVF International Conference on Computer Vision , pages=

  6. [14]

    Seminal Graphics Papers: Pushing the Boundaries, Volume 2 , pages=

    SMPL: A skinned multi-person linear model , author=. Seminal Graphics Papers: Pushing the Boundaries, Volume 2 , pages=

  7. [15]

    ACM Transactions on Multimedia Computing, Communications, and Applications (TOMM) , volume=

    Music2dance: Dancenet for music-driven dance generation , author=. ACM Transactions on Multimedia Computing, Communications, and Applications (TOMM) , volume=. 2022 , publisher=

    2022

  8. [16]

    Proceedings of the IEEE/CVF International Conference on Computer Vision , pages=

    Finedance: A fine-grained choreography dataset for 3d full body dance generation , author=. Proceedings of the IEEE/CVF International Conference on Computer Vision , pages=

  9. [17]

    Proceedings of the 30th ACM International Conference on Multimedia , pages=

    Groupdancer: Music to multi-people dance synthesis with style collaboration , author=. Proceedings of the 30th ACM International Conference on Multimedia , pages=

  10. [18]

    , author=

    AIST Dance Video Database: Multi-Genre, Multi-Dancer, and Multi-Camera Database for Dance Information Processing. , author=. ISMIR , volume=

  11. [19]

    Communications of the ACM , volume=

    Generative adversarial networks , author=. Communications of the ACM , volume=. 2020 , publisher=

    2020

  12. [20]

    YouTube , howpublished =

  13. [21]

    2024 , note =

    Blender Foundation , title =. 2024 , note =

    2024

  14. [22]

    Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages=

    On the continuity of rotation representations in neural networks , author=. Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages=

  15. [23]

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

    Revisiting temporal modeling for clip-based image-to-video knowledge transferring , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

  16. [24]

    arXiv preprint arXiv:2306.00107 , year=

    Mert: Acoustic music understanding model with large-scale self-supervised training , author=. arXiv preprint arXiv:2306.00107 , year=

    arXiv

  17. [25]

    Proceedings of the 2024 International Conference on Multimedia Retrieval , pages=

    BeatDance: A Beat-Based Model-Agnostic Contrastive Learning Framework for Music-Dance Retrieval , author=. Proceedings of the 2024 International Conference on Multimedia Retrieval , pages=

    2024

  18. [26]

    Proceedings of the IEEE/CVF International Conference on Computer Vision , pages=

    Actformer: A gan-based transformer towards general action-conditioned 3d human motion generation , author=. Proceedings of the IEEE/CVF International Conference on Computer Vision , pages=

  19. [27]

    Advances in neural information processing systems , volume=

    Scheduled sampling for sequence prediction with recurrent neural networks , author=. Advances in neural information processing systems , volume=

  20. [28]

    arXiv preprint arXiv:2009.08553 , year=

    Generation-augmented retrieval for open-domain question answering , author=. arXiv preprint arXiv:2009.08553 , year=

    2009 arXiv

  21. [29]

    , author=

    FMDistance: A Fast and Effective Distance Function for Motion Capture Data. , author=. Eurographics (Short Papers) , volume=

  22. [30]

    ACM SIGGRAPH 2005 Papers , pages=

    Efficient content-based retrieval of motion capture data , author=. ACM SIGGRAPH 2005 Papers , pages=

    2005

  23. [31]

    arXiv preprint arXiv:2005.00341 , year=

    Jukebox: A generative model for music , author=. arXiv preprint arXiv:2005.00341 , year=

    2005 arXiv

  24. [32]

    , author=

    librosa: Audio and music signal analysis in python. , author=. SciPy , pages=

  25. [33]

    IEEE Transactions on Multimedia , volume=

    Deepdance: music-to-dance motion choreography with adversarial learning , author=. IEEE Transactions on Multimedia , volume=. 2020 , publisher=

    2020

  26. [34]

    Proceedings of the 31st ACM International Conference on Multimedia , pages=

    Dance with you: The diversity controllable dancer generation via diffusion models , author=. Proceedings of the 31st ACM International Conference on Multimedia , pages=

  27. [35]

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

    Bidirectional Autoregessive Diffusion Model for Dance Generation , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

  28. [36]

    2018 10th international conference on virtual worlds and games for serious applications (VS-Games) , pages=

    An embodied learning game using kinect and labanotation for analysis and visualization of dance kinesiology , author=. 2018 10th international conference on virtual worlds and games for serious applications (VS-Games) , pages=. 2018 , organization=

    2018

  29. [37]

    Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages=

    Expressive body capture: 3d hands, face, and body from a single image , author=. Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages=

  30. [38]

    arXiv preprint arXiv:2201.02610 , year=

    Embodied hands: Modeling and capturing hands and bodies together , author=. arXiv preprint arXiv:2201.02610 , year=

    arXiv

  31. [39]

    Proceedings of the AAAI Conference on Artificial Intelligence , volume=

    Danceformer: Music conditioned 3d dance generation with parametric motion transformer , author=. Proceedings of the AAAI Conference on Artificial Intelligence , volume=

  32. [40]

    ICASSP 2023-2023 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) , pages=

    GTN-Bailando: Genre consistent long-term 3d dance generation based on pre-trained genre token network , author=. ICASSP 2023-2023 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) , pages=. 2023 , organization=

    2023

  33. [41]

    2024 , howpublished =

    Blender Foundation , title =. 2024 , howpublished =

    2024

  34. [42]

    Retrieved , volume=

    YouTube , author=. Retrieved , volume=

  35. [43]

    Advances in Neural Information Processing Systems , year=

    Attention is all you need , author=. Advances in Neural Information Processing Systems , year=

  36. [44]

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

    Mvbench: A comprehensive multi-modal video understanding benchmark , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

  37. [45]

    Phenomenology and the Cognitive Sciences , volume=

    Perceiving subjectivity in bodily movement: The case of dancers , author=. Phenomenology and the Cognitive Sciences , volume=. 2009 , publisher=

    2009

  38. [46]

    arXiv preprint arXiv:1811.00818 , year=

    Listen to dance: Music-driven choreography generation using autoregressive encoder-decoder network , author=. arXiv preprint arXiv:1811.00818 , year=

    arXiv

  39. [47]

    Proceedings of the 28th ACM International Conference on Multimedia , pages=

    Self-supervised dance video synthesis conditioned on music , author=. Proceedings of the 28th ACM International Conference on Multimedia , pages=

  40. [48]

    Proceedings of the 26th ACM international conference on Multimedia , pages=

    Dance with melody: An lstm-autoencoder approach to music-oriented dance synthesis , author=. Proceedings of the 26th ACM international conference on Multimedia , pages=

  41. [49]

    ACM Transactions on Graphics (TOG) , volume=

    Choreomaster: choreography-oriented music-driven dance synthesis , author=. ACM Transactions on Graphics (TOG) , volume=. 2021 , publisher=

    2021

  42. [50]

    arXiv preprint arXiv:2006.06119 , year=

    Dance revolution: Long-term dance generation with music via curriculum learning , author=. arXiv preprint arXiv:2006.06119 , year=

    2006 arXiv

  43. [51]

    Proceedings of the 28th ACM International Conference on Multimedia , pages=

    Action2motion: Conditioned generation of 3d human motions , author=. Proceedings of the 28th ACM International Conference on Multimedia , pages=

  44. [52]

    Proceedings of the IEEE/CVF International Conference on Computer Vision , pages=

    Action-conditioned 3d human motion synthesis with transformer vae , author=. Proceedings of the IEEE/CVF International Conference on Computer Vision , pages=

  45. [53]

    arXiv preprint arXiv:2404.11375 , year=

    Text-controlled Motion Mamba: Text-Instructed Temporal Grounding of Human Motion , author=. arXiv preprint arXiv:2404.11375 , year=

  46. [54]

    arXiv preprint arXiv:2407.10061 , year=

    Infinimotion: Mamba boosts memory in transformer for arbitrary long motion generation , author=. arXiv preprint arXiv:2407.10061 , year=

    arXiv

  47. [55]

    European Conference on Computer Vision , pages=

    Tm2t: Stochastic and tokenized modeling for the reciprocal generation of 3d human motions and texts , author=. European Conference on Computer Vision , pages=. 2022 , organization=

    2022

  48. [56]

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

    Generating diverse and natural 3d human motions from text , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

  49. [57]

    IEEE Transactions on Pattern Analysis and Machine Intelligence , year=

    Motiondiffuse: Text-driven human motion generation with diffusion model , author=. IEEE Transactions on Pattern Analysis and Machine Intelligence , year=

  50. [58]

    Proceedings of the IEEE/CVF International Conference on Computer Vision , pages=

    Make-an-animation: Large-scale text-conditional 3D human motion generation , author=. Proceedings of the IEEE/CVF International Conference on Computer Vision , pages=

  51. [59]

    Proceedings of the AAAI Conference on Artificial Intelligence , volume=

    Everything2Motion: Synchronizing Diverse Inputs via a Unified Framework for Human Motion Synthesis , author=. Proceedings of the AAAI Conference on Artificial Intelligence , volume=

  52. [60]

    ICASSP 2024-2024 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) , pages=

    Exploring Multi-Modal Control in Music-Driven Dance Generation , author=. ICASSP 2024-2024 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) , pages=. 2024 , organization=

    2024

  53. [61]

    ICASSP 2022-2022 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) , pages=

    Genre-conditioned long-term 3d dance generation driven by music , author=. ICASSP 2022-2022 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) , pages=. 2022 , organization=

    2022

  54. [62]

    arXiv preprint arXiv:2403.09407 , year=

    LM2D: Lyrics-and Music-Driven Dance Synthesis , author=. arXiv preprint arXiv:2403.09407 , year=

    arXiv

  55. [63]

    Proceedings of the IEEE/CVF international conference on computer vision , pages=

    AMASS: Archive of motion capture as surface shapes , author=. Proceedings of the IEEE/CVF international conference on computer vision , pages=

  56. [64]

    Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages=

    Learning 3d human dynamics from video , author=. Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages=

  57. [65]

    Proceedings of the European conference on computer vision (ECCV) , pages=

    Recovering accurate 3d human pose in the wild using imus and a moving camera , author=. Proceedings of the European conference on computer vision (ECCV) , pages=

  58. [66]

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

    Human3. 6m: Large scale datasets and predictive methods for 3d human sensing in natural environments , author=. IEEE transactions on pattern analysis and machine intelligence , volume=. 2013 , publisher=

    2013

  59. [67]

    Proceedings of the IEEE Conference on computer Vision and Pattern Recognition , pages=

    2d human pose estimation: New benchmark and state of the art analysis , author=. Proceedings of the IEEE Conference on computer Vision and Pattern Recognition , pages=

  60. [68]

    Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages=

    Deep high-resolution representation learning for human pose estimation , author=. Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages=

  61. [69]

    International conference on machine learning , pages=

    Learning transferable visual models from natural language supervision , author=. International conference on machine learning , pages=. 2021 , organization=

    2021

  62. [70]

    Neural computation , volume=

    Adaptive mixtures of local experts , author=. Neural computation , volume=. 1991 , publisher=

    1991

  63. [71]

    Computers & Graphics , volume=

    Learning to dance: A graph convolutional adversarial network to generate realistic dance motions from audio , author=. Computers & Graphics , volume=. 2021 , publisher=

    2021

  64. [72]

    Proceedings of the IEEE/CVF International Conference on Computer Vision , pages=

    Convolutional sequence generation for skeleton-based action synthesis , author=. Proceedings of the IEEE/CVF International Conference on Computer Vision , pages=

  65. [73]

    arXiv preprint arXiv:2412.19123 , year=

    CoheDancers: Enhancing Interactive Group Dance Generation through Music-Driven Coherence Decomposition , author=. arXiv preprint arXiv:2412.19123 , year=

    arXiv

  66. [74]

    arXiv preprint arXiv:2410.20389 , year=

    Lodge++: High-quality and Long Dance Generation with Vivid Choreography Patterns , author=. arXiv preprint arXiv:2410.20389 , year=

    arXiv

  67. [75]

    Proceedings of the 28th ACM International Conference on Multimedia , pages=

    Choreonet: Towards music to dance synthesis with choreographic action unit , author=. Proceedings of the 28th ACM International Conference on Multimedia , pages=

  68. [76]

    IEEE Transactions on Visualization and Computer Graphics , year=

    Rhythm is a dancer: Music-driven motion synthesis with global structure , author=. IEEE Transactions on Visualization and Computer Graphics , year=

  69. [77]

    The Thirty-eighth Annual Conference on Neural Information Processing Systems , year=

    Mambatalk: Efficient holistic gesture synthesis with selective state space models , author=. The Thirty-eighth Annual Conference on Neural Information Processing Systems , year=

  70. [78]

    arXiv preprint arXiv:2403.15569 , year=

    Music to Dance as Language Translation using Sequence Models , author=. arXiv preprint arXiv:2403.15569 , year=

    arXiv

  71. [79]

    Advances in neural information processing systems , volume=

    Dancing to music , author=. Advances in neural information processing systems , volume=

  72. [80]

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

    Vbench: Comprehensive benchmark suite for video generative models , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

  73. [81]

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

    Evalcrafter: Benchmarking and evaluating large video generation models , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

  74. [82]

    arXiv preprint arXiv:2501.01108 , year=

    Muq: Self-supervised music representation learning with mel residual vector quantization , author=. arXiv preprint arXiv:2501.01108 , year=

    arXiv

  75. [83]

    arXiv preprint arXiv:2312.00752 , year=

    Mamba: Linear-time sequence modeling with selective state spaces , author=. arXiv preprint arXiv:2312.00752 , year=

    arXiv

  76. [84]

    arXiv preprint arXiv:2309.15505 , year=

    Finite scalar quantization: Vq-vae made simple , author=. arXiv preprint arXiv:2309.15505 , year=

    arXiv

  77. [85]

    Advances in neural information processing systems , volume=

    Neural discrete representation learning , author=. Advances in neural information processing systems , volume=

  78. [86]

    arXiv preprint arXiv:2004.05150 , year=

    Longformer: The long-document transformer , author=. arXiv preprint arXiv:2004.05150 , year=

    2004 arXiv

  79. [87]

    Proceedings of the IEEE/CVF International Conference on Computer Vision , pages=

    Video background music generation: Dataset, method and evaluation , author=. Proceedings of the IEEE/CVF International Conference on Computer Vision , pages=

  80. [88]

    2018 , publisher=

    Improving language understanding by generative pre-training , author=. 2018 , publisher=

    2018

  81. [89]

    Proceedings of the 32nd ACM International Conference on Multimedia , pages=

    MambaGesture: Enhancing Co-Speech Gesture Generation with Mamba and Disentangled Multi-Modality Fusion , author=. Proceedings of the 32nd ACM International Conference on Multimedia , pages=

  82. [90]

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

    Wham: Reconstructing world-grounded humans with accurate 3d motion , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

  83. [91]

    International Conference on Learning Representations (ICLR) , year=

    Outrageously large neural networks: The sparsely-gated mixture-of-experts layer , author=. International Conference on Learning Representations (ICLR) , year=

  84. [92]

    Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics (ACL) , pages=

    Don't stop pretraining: Adapt language models to domains and tasks , author=. Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics (ACL) , pages=

  85. [93]

    International conference on machine learning , pages=

    Weakly-supervised disentanglement without compromises , author=. International conference on machine learning , pages=. 2020 , organization=

    2020

  86. [94]

    Le and Geoffrey E

    Noam Shazeer and Azalia Mirhoseini and Krzysztof Maziarz and Andy Davis and Quoc V. Le and Geoffrey E. Hinton and Jeff Dean , title =. arXiv preprint arXiv:1701.06538 , year =

    arXiv

  87. [95]

    arXiv preprint arXiv:2111.00396 , year=

    Efficiently modeling long sequences with structured state spaces , author=. arXiv preprint arXiv:2111.00396 , year=

    arXiv

  88. [96]

    Advances in Neural Information Processing Systems (NeurIPS) , year=

    Neural Discrete Representation Learning , author=. Advances in Neural Information Processing Systems (NeurIPS) , year=

  89. [97]

    Neural Computing and Applications , volume=

    Choreography cGAN: generating dances with music beats using conditional generative adversarial networks , author=. Neural Computing and Applications , volume=. 2021 , publisher=

    2021

  90. [98]

    arXiv preprint arXiv:2210.02747 , year=

    Flow matching for generative modeling , author=. arXiv preprint arXiv:2210.02747 , year=

    arXiv

  91. [99]

    arXiv preprint arXiv:2505.13447 , year=

    Mean flows for one-step generative modeling , author=. arXiv preprint arXiv:2505.13447 , year=

    arXiv

  92. [100]

    arXiv preprint arXiv:2505.17543 , year=

    Megadance: Mixture-of-experts architecture for genre-aware 3d dance generation , author=. arXiv preprint arXiv:2505.17543 , year=

  93. [101]

    arXiv preprint arXiv:2502.18309 , year=

    GCDance: Genre-Controlled 3D Full Body Dance Generation Driven By Music , author=. arXiv preprint arXiv:2502.18309 , year=

    arXiv

  94. [102]

    Advances in neural information processing systems , volume=

    Denoising diffusion probabilistic models , author=. Advances in neural information processing systems , volume=

  95. [103]

    Forty-first international conference on machine learning , year=

    Scaling rectified flow transformers for high-resolution image synthesis , author=. Forty-first international conference on machine learning , year=

  96. [104]

    arXiv preprint arXiv:2504.05304 , year=

    Gaussian mixture flow matching models , author=. arXiv preprint arXiv:2504.05304 , year=

    arXiv

  97. [105]

    arXiv preprint arXiv:2410.05954 , year=

    Pyramidal flow matching for efficient video generative modeling , author=. arXiv preprint arXiv:2410.05954 , year=

  98. [106]

    Proceedings of the IEEE/CVF International Conference on Computer Vision , pages=

    Efficient video prediction via sparsely conditioned flow matching , author=. Proceedings of the IEEE/CVF International Conference on Computer Vision , pages=

  99. [107]

    arXiv preprint arXiv:2209.03003 , year=

    Flow straight and fast: Learning to generate and transfer data with rectified flow , author=. arXiv preprint arXiv:2209.03003 , year=

    arXiv

  100. [108]

    arXiv preprint arXiv:2407.02398 , year=

    Consistency flow matching: Defining straight flows with velocity consistency , author=. arXiv preprint arXiv:2407.02398 , year=

    arXiv

  101. [109]

    Advances in neural information processing systems , volume=

    Voicebox: Text-guided multilingual universal speech generation at scale , author=. Advances in neural information processing systems , volume=

  102. [110]

    arXiv preprint arXiv:2312.15821 , year=

    Audiobox: Unified audio generation with natural language prompts , author=. arXiv preprint arXiv:2312.15821 , year=

    arXiv

  103. [111]

    ICASSP 2024-2024 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) , pages=

    Matcha-TTS: A fast TTS architecture with conditional flow matching , author=. ICASSP 2024-2024 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) , pages=. 2024 , organization=

    2024

  104. [112]

    ICASSP 2024-2024 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) , pages=

    Voiceflow: Efficient text-to-speech with rectified flow matching , author=. ICASSP 2024-2024 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) , pages=. 2024 , organization=

    2024

  105. [113]

    arXiv preprint arXiv:2312.08895 , year=

    Motion flow matching for human motion synthesis and editing , author=. arXiv preprint arXiv:2312.08895 , year=

    arXiv

  106. [114]

    arXiv preprint arXiv:2412.01064 , year=

    Float: Generative motion latent flow matching for audio-driven talking portrait , author=. arXiv preprint arXiv:2412.01064 , year=

  107. [115]

    1988 , publisher=

    The moment of movement: Dance improvisation , author=. 1988 , publisher=

    1988

  108. [116]

    Dance Research Journal , volume=

    Dance studies/cultural studies , author=. Dance Research Journal , volume=. 2009 , publisher=

    2009

  109. [117]

    arXiv preprint arXiv:1702.01992 , year=

    Gated multimodal units for information fusion , author=. arXiv preprint arXiv:1702.01992 , year=

    arXiv

  110. [118]

    Proceedings of the AAAI conference on artificial intelligence , volume=

    Film: Visual reasoning with a general conditioning layer , author=. Proceedings of the AAAI conference on artificial intelligence , volume=

  111. [119]

    Proceedings of the 26th international conference on world wide web , pages=

    Neural collaborative filtering , author=. Proceedings of the 26th international conference on world wide web , pages=

  112. [120]

    arXiv preprint arXiv:2505.14222 , year=

    MatchDance: Collaborative Mamba-Transformer Architecture Matching for High-Quality 3D Dance Synthesis , author=. arXiv preprint arXiv:2505.14222 , year=

  113. [121]

    arXiv preprint arXiv:2506.18866 , year=

    OmniAvatar: Efficient Audio-Driven Avatar Video Generation with Adaptive Body Animation , author=. arXiv preprint arXiv:2506.18866 , year=

    arXiv

  114. [122]

    Neue Methoden zur approximativen Integration der Differentialgleichungen einer unabh

    Heun, Karl and others , journal=. Neue Methoden zur approximativen Integration der Differentialgleichungen einer unabh

  115. [123]

    The Fourth Blogpost Track at ICLR 2025 , year=

    Diffusion models and gaussian flow matching: Two sides of the same coin , author=. The Fourth Blogpost Track at ICLR 2025 , year=

    2025

  116. [124]

    2025 , booktitle=

    Mu, Yuxuan and Ling, Hung Yu and Shi, Yi and Baira Ojeda, Ismael and Xi, Pengcheng and Shu, Chang and Zinno, Fabio and Peng, Xue Bin , title=. 2025 , booktitle=

    2025

  117. [125]

    arXiv preprint arXiv:2503.18886 , year=

    Cfg-zero*: Improved classifier-free guidance for flow matching models , author=. arXiv preprint arXiv:2503.18886 , year=

    arXiv

  118. [126]

    arXiv preprint arXiv:2501.15420 , year=

    Visual generation without guidance , author=. arXiv preprint arXiv:2501.15420 , year=

    arXiv

  119. [127]

    1993 , publisher=

    Solving ordinary differential equations I: Nonstiff problems , author=. 1993 , publisher=

    1993

  120. [128]

    arXiv preprint arXiv:2410.10306 , year=

    Animate-x: Universal character image animation with enhanced motion representation , author=. arXiv preprint arXiv:2410.10306 , year=

    arXiv

  121. [129]

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

    Animate anyone: Consistent and controllable image-to-video synthesis for character animation , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

  122. [130]

    arXiv preprint arXiv:2509.14055 , year=

    Wan-Animate: Unified Character Animation and Replacement with Holistic Replication , author=. arXiv preprint arXiv:2509.14055 , year=

  123. [131]

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

    Synctalk: The devil is in the synchronization for talking head synthesis , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

  124. [132]

    arXiv preprint arXiv:2506.14742 , year=

    SyncTalk++: High-Fidelity and Efficient Synchronized Talking Heads Synthesis Using Gaussian Splatting , author=. arXiv preprint arXiv:2506.14742 , year=

    arXiv

  125. [133]

    Proceedings of the IEEE/CVF international conference on computer vision , pages=

    Emotalk: Speech-driven emotional disentanglement for 3d face animation , author=. Proceedings of the IEEE/CVF international conference on computer vision , pages=

  126. [134]

    Proceedings of the 31st ACM International Conference on Multimedia , pages=

    Selftalk: A self-supervised commutative training diagram to comprehend 3d talking faces , author=. Proceedings of the 31st ACM International Conference on Multimedia , pages=

  127. [135]

    arXiv preprint arXiv:2505.21448 , year=

    Omnisync: Towards universal lip synchronization via diffusion transformers , author=. arXiv preprint arXiv:2505.21448 , year=

  128. [136]

    arXiv preprint arXiv:2502.17414 , year=

    X-dancer: Expressive music to human dance video generation , author=. arXiv preprint arXiv:2502.17414 , year=

    arXiv

  129. [137]

    2025 IEEE/CVF Winter Conference on Applications of Computer Vision (WACV) , pages=

    Dance any beat: Blending beats with visuals in dance video generation , author=. 2025 IEEE/CVF Winter Conference on Applications of Computer Vision (WACV) , pages=. 2025 , organization=

    2025

  130. [138]

    Advances in neural information processing systems , volume=

    Vitpose: Simple vision transformer baselines for human pose estimation , author=. Advances in neural information processing systems , volume=

  131. [139]

    arXiv preprint arXiv:2207.12598 , year=

    Classifier-free diffusion guidance , author=. arXiv preprint arXiv:2207.12598 , year=

    arXiv

  132. [140]

    , author=

    Lora: Low-rank adaptation of large language models. , author=. ICLR , volume=

  133. [141]

    arXiv preprint arXiv:2410.07718 , year=

    Hallo2: Long-duration and high-resolution audio-driven portrait image animation , author=. arXiv preprint arXiv:2410.07718 , year=

    arXiv

  134. [142]

    International Conference on Learning Representations (ICLR) , year=

    Denoising Diffusion Implicit Models , author=. International Conference on Learning Representations (ICLR) , year=

  135. [143]

    Proceedings of the 32nd ACM International Conference on Multimedia , pages=

    Transnet v2: An effective deep network architecture for fast shot transition detection , author=. Proceedings of the 32nd ACM International Conference on Multimedia , pages=

  136. [144]

    Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages=

    Vibe: Video inference for human body pose and shape estimation , author=. Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages=

  137. [145]

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

    Magicanimate: Temporally consistent human image animation using diffusion model , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

  138. [146]

    arXiv preprint arXiv:2511.21029 , year=

    FlowerDance: MeanFlow for Efficient and Refined 3D Dance Generation , author=. arXiv preprint arXiv:2511.21029 , year=

    arXiv

  139. [147]

    Proceedings of the IEEE/CVF International Conference on Computer Vision , pages=

    MDD: A Dataset for Text-and-Music Conditioned Duet Dance Generation , author=. Proceedings of the IEEE/CVF International Conference on Computer Vision , pages=

  140. [148]

    ACM Transactions on Graphics (TOG) , volume=

    Interactive motion generation from examples , author=. ACM Transactions on Graphics (TOG) , volume=. 2002 , publisher=

    2002

  141. [149]

    Proceedings of the 12th ACM SIGGRAPH Conference on Motion, Interaction and Games , pages=

    Spatial motion doodles: Sketching animation in vr using hand gestures and laban motion analysis , author=. Proceedings of the 12th ACM SIGGRAPH Conference on Motion, Interaction and Games , pages=

  142. [150]

    ACM Transactions on Graphics (ToG) , volume=

    Motion doodles: an interface for sketching character motion , author=. ACM Transactions on Graphics (ToG) , volume=. 2004 , publisher=

    2004

  143. [151]

    Proceedings of the 2024 CHI Conference on Human Factors in Computing Systems , pages=

    TimeTunnel: Integrating Spatial and Temporal Motion Editing for Character Animation in Virtual Reality , author=. Proceedings of the 2024 CHI Conference on Human Factors in Computing Systems , pages=

    2024

  144. [152]

    Proceedings of the 2024 ACM Designing Interactive Systems Conference , pages=

    Dancegen: Supporting choreography ideation and prototyping with generative ai , author=. Proceedings of the 2024 ACM Designing Interactive Systems Conference , pages=

    2024

  145. [153]

    Proceedings of the 2025 CHI Conference on Human Factors in Computing Systems , pages=

    ChoreoCraft: In-situ Crafting of Choreography in Virtual Reality through Creativity Support Tool , author=. Proceedings of the 2025 CHI Conference on Human Factors in Computing Systems , pages=

    2025

  146. [154]

    Pragmatics & Cognition , volume=

    Choreographic cognition: The time-course and phenomenology of creating a dance , author=. Pragmatics & Cognition , volume=. 2003 , publisher=

    2003

  147. [155]

    Research in dance education , volume=

    Teaching choreography in higher education: A process continuum model , author=. Research in dance education , volume=. 2004 , publisher=

    2004

  148. [156]

    Dance Research , volume=

    Thinking strategically about dance making: an analysis of the structuring stage and the strategies choreographers use for varying dance works , author=. Dance Research , volume=. 2020 , publisher=

    2020

  149. [157]

    2014 , publisher=

    Dance composition: A practical guide to creative success in dance making , author=. 2014 , publisher=

    2014

  150. [158]

    1982 , publisher=

    The intimate act of choreography , author=. 1982 , publisher=

    1982

  151. [159]

    ACM Transactions on Graphics (TOG) , volume=

    Performance-based control interface for character animation , author=. ACM Transactions on Graphics (TOG) , volume=. 2009 , publisher=

    2009

  152. [160]

    Computer Animation and Virtual Worlds , volume=

    Performance-driven motion choreographing with accelerometers , author=. Computer Animation and Virtual Worlds , volume=. 2009 , publisher=

    2009

  153. [161]

    The Visual Computer , volume=

    Composition of multiple figure sequences for dance and animation , author=. The Visual Computer , volume=. 1991 , publisher=

    1991

  154. [162]

    ACM Transactions on Graphics (ToG) , volume=

    SketchiMo: sketch-based motion editing for articulated characters , author=. ACM Transactions on Graphics (ToG) , volume=. 2016 , publisher=

    2016

  155. [163]

    Proceedings of the 2015 ACM SIGCHI Conference on Creativity and Cognition , pages=

    Drawing apprentice: An enactive co-creative agent for artistic collaboration , author=. Proceedings of the 2015 ACM SIGCHI Conference on Creativity and Cognition , pages=

    2015

  156. [164]

    National Science Review , volume=

    A survey on multimodal large language models , author=. National Science Review , volume=. 2024 , publisher=

    2024

  157. [165]

    arXiv preprint arXiv:2312.11805 , year=

    Gemini: a family of highly capable multimodal models , author=. arXiv preprint arXiv:2312.11805 , year=

    arXiv

  158. [166]

    arXiv preprint arXiv:2209.14916 , year=

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

    arXiv

  159. [167]

    Proceedings of the AAAI Conference on Artificial Intelligence , volume=

    Flame: Free-form language-based motion synthesis & editing , author=. Proceedings of the AAAI Conference on Artificial Intelligence , volume=

  160. [168]

    arXiv preprint arXiv:2008.08171 , year=

    Learning to generate diverse dance motions with transformer , author=. arXiv preprint arXiv:2008.08171 , year=

    2008 arXiv

  161. [169]

    Journal of graphics tools , volume=

    Practical parameterization of rotations using the exponential map , author=. Journal of graphics tools , volume=. 1998 , publisher=

    1998

  162. [170]

    , title =

    Yamane, Katsu and Hodgins, Jessica K. , title =. IEEE Transactions on Robotics , year =

  163. [171]

    Proceedings of the 26th Annual ACM Symposium on User Interface Software and Technology , pages =

    Anderson, Fraser and Grossman, Tovi and Matejka, Justin and Fitzmaurice, George , title =. Proceedings of the 26th Annual ACM Symposium on User Interface Software and Technology , pages =. 2013 , isbn =. doi:10.1145/2501988.2502045 , abstract =

    2013

  164. [172]

    WARPY: Sketching Environment-Aware 3D Curves in Mobile Augmented Reality , year=

    Alghofaili, Rawan and Nguyen, Cuong and Krs, Vojtĕch and Carr, Nathan and Mĕch, Radomír and Yu, Lap-Fai , booktitle=. WARPY: Sketching Environment-Aware 3D Curves in Mobile Augmented Reality , year=

  165. [173]

    ACM SIGGRAPH 2024 Conference Papers , articleno =

    Kim, Minyoung and Alghofaili, Rawan and Li, Changyang and Yu, Lap-Fai , title =. ACM SIGGRAPH 2024 Conference Papers , articleno =. 2024 , isbn =. doi:10.1145/3641519.3657397 , abstract =

    2024

  166. [174]

    ACM Trans

    Ye, Hui and Kwan, Kin Chung and Su, Wanchao and Fu, Hongbo , title =. ACM Trans. Graph. , month = aug, articleno =. 2020 , issue_date =. doi:10.1145/3386569.3392404 , abstract =

    2020

  167. [175]

    IEEE Transactions on Visualization and Computer Graphics , number=

    Keyframe control of music-driven 3d dance generation , author=. IEEE Transactions on Visualization and Computer Graphics , number=. 2023 , publisher=

    2023

  168. [176]

    Empirical Musicology Review , volume=

    Pleasurable and intersubjectively embodied experiences of electronic dance music , author=. Empirical Musicology Review , volume=. 2017 , publisher=

    2017

  169. [177]

    Plos one , volume=

    Music we move to: Spotify audio features and reasons for listening , author=. Plos one , volume=. 2022 , publisher=

    2022

  170. [178]

    2012 , publisher=

    Basic principles of classical ballet: Russian ballet technique , author=. 2012 , publisher=

    2012

  171. [179]

    Proceedings of the IEEE/CVF International Conference on Computer Vision , pages=

    Music-aligned holistic 3d dance generation via hierarchical motion modeling , author=. Proceedings of the IEEE/CVF International Conference on Computer Vision , pages=

  172. [180]

    arXiv preprint arXiv:2509.24099 , year=

    Unified multi-modal interactive & reactive 3d motion generation via rectified flow , author=. arXiv preprint arXiv:2509.24099 , year=

  173. [181]

    ACM Transactions on Graphics (TOG) , volume=

    Control Operators for Interactive Character Animation , author=. ACM Transactions on Graphics (TOG) , volume=. 2025 , publisher=

    2025

  174. [182]

    2026 IEEE Conference on Virtual Reality and 3D User Interfaces (VR) , pages=

    Personalized Dance Synthesis Based on Physical and Cognitive Intensities , author=. 2026 IEEE Conference on Virtual Reality and 3D User Interfaces (VR) , pages=. 2026 , organization=

    2026

  175. [183]

    arXiv preprint arXiv:2606.30019 , year=

    OmniDance: Multimodal Driven Dance Video Generation with Large-scale Internet Data , author=. arXiv preprint arXiv:2606.30019 , year=

    arXiv

  176. [184]

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

    PromptHMR: Promptable Human Mesh Recovery , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

  177. [185]

    Proceedings of the IEEE/CVF International Conference on Computer Vision , pages=

    EgoMusic-driven Human Dance Motion Estimation with Skeleton Mamba , author=. Proceedings of the IEEE/CVF International Conference on Computer Vision , pages=

  178. [186]

    Proceedings of the IEEE/CVF International Conference on Computer Vision , pages=

    Align Your Rhythm: Generating Highly Aligned Dance Poses with Gating-Enhanced Rhythm-Aware Feature Representation , author=. Proceedings of the IEEE/CVF International Conference on Computer Vision , pages=

  179. [187]

    arXiv preprint arXiv:2506.02661 , year=

    MotionRAG-Diff: A Retrieval-Augmented Diffusion Framework for Long-Term Music-to-Dance Generation , author=. arXiv preprint arXiv:2506.02661 , year=

    arXiv

  180. [188]

    arXiv preprint arXiv:2504.04634 , year=

    DanceMosaic: High-Fidelity Dance Generation with Multimodal Editability , author=. arXiv preprint arXiv:2504.04634 , year=

  181. [189]

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

    POPDG: Popular 3D Dance Generation with PopDanceSet , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

  182. [190]

    Proceedings of the IEEE/CVF International Conference on Computer Vision , pages=

    TMR: Text-to-Motion Retrieval Using Contrastive 3D Human Motion Synthesis , author=. Proceedings of the IEEE/CVF International Conference on Computer Vision , pages=

  183. [191]

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

    Tri-Modal Motion Retrieval by Learning a Joint Embedding Space , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

  184. [192]

    Proceedings of the Special Interest Group on Computer Graphics and Interactive Techniques Conference Conference Papers , pages=

    Mace-dance: Motion-appearance cascaded experts for music-driven dance video generation , author=. Proceedings of the Special Interest Group on Computer Graphics and Interactive Techniques Conference Conference Papers , pages=

  185. [193]

    arXiv preprint arXiv:2604.04395 , year=

    BiTDiff: Fine-Grained 3D Conducting Motion Generation via BiMamba-Transformer Diffusion , author=. arXiv preprint arXiv:2604.04395 , year=

    arXiv

  186. [194]

    arXiv preprint arXiv:2603.27314 , year=

    Tokendance: Token-to-token music-to-dance generation with bidirectional mamba , author=. arXiv preprint arXiv:2603.27314 , year=

  187. [195]

    2026 , eprint=

    StreamTalk: Streaming Co-Speech Gesture Generation with Key-Pose Anchoring , author=. 2026 , eprint=

    2026

  188. [196]

    arXiv preprint arXiv:2605.06064 , year=

    PersonaGesture: Single-Reference Co-Speech Gesture Personalization for Unseen Speakers , author=. arXiv preprint arXiv:2605.06064 , year=

    arXiv

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.