Pith. sign in

REVIEW 4 major objections 4 minor 92 references

ETHead: Generating Expressive 3D Facial Animation and Head Movement from Speech

T0 review · 4 major / 4 minor · reviewed 2026-08-05 · deepseek-v4-flash

Pith's one-line read ETHead claims that an encoder pre-trained on 2D talking video drives emotionally expressive 3D facial and head animation that beats state-of-the-art methods, and transfers to other talking-head frameworks.

desk verdict ETHead is a serious systems contribution with a new audio-visual speech encoder, but its headline superiority claim is compromised by training and evaluating on the same SMIRK pseudo-ground truth. read the letter →

arxiv 2608.01605 v1 pith:MEKTQJXH submitted 2026-08-03 cs.GR

classification cs.GR
keywords 3Dtalkingheadspeech-drivenfacialanimationexpressivemotiongenerationposesynthesisself-distillationemotion-awaremaskingspeechrepresentationlearningdiffusionmodel
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

The paper sets out to solve a data problem: expressive 3D talking-head models are starved for high-fidelity 3D emotional speech data, so they cannot learn the complex facial and head motions that carry emotion. ETHead attacks this by pre-training a speech encoder on hundreds of hours of ordinary 2D talking-head video, using a student-teacher self-distillation scheme in which an emotion-aware masking mechanism forces the model to reconstruct the audio-visual dynamics of emotionally salient moments. The encoder then feeds a diffusion-based 3D generator in two ways: it injects motion-aligned cues into the speech conditioning, and it supervises the output through a joint speech-motion latent space. The paper reports that this combination beats three state-of-the-art baselines on standard emotion datasets, improves head-pose realism and beat alignment, runs at 20.48 frames per second, and that the encoder can be bolted onto other 3D talking-head pipelines to improve them. If right, it shows that the 3D-data bottleneck can be bypassed by transferring motion knowledge from abundant 2D video.

What carries the argument

The load-bearing object is the motion-aligned speech encoder, a student-teacher self-distillation network adapted from DINO-style masked modeling. The student sees audio and visual tokens that have been partially masked, with masking probability raised at emotionally salient moments identified by normalized pitch (F0) deviation, while the teacher sees the full, unmasked clip; the student must reconstruct the teacher's fused audio-visual representation at masked positions (regression loss) and match its soft class distribution (cross-entropy loss). Stochastic modality dropout lets the speech branch run alone at inference. The masked-reconstruction asymmetry is what forces the audio branch to

What would settle it

Run the identical ETHead pipeline and evaluate it on high-frame-rate 4D motion-capture ground truth (multi-view or depth capture, not monocular reconstruction) for the same actors and utterances; if the LVE/EVE/FDD/BA/FID gaps over DiffPoseTalk shrink to noise once the reconstruction pipeline is removed, the central claim is refuted. A cheaper check: split the RAVDESS test frames by SMIRK reconstruction confidence and verify that emotion-alignment gains concentrate in high-confidence frames; if the gains are absent or inverted where reconstruction is noisy, they are artifacts of the pseudo-gro

Watch

Extended reading notes

Core claim

ETHead's central claim is that speech carries enough information about expressive facial and head motion, provided it is decoded with a representation that has seen visual dynamics, to generate emotionally coherent 3D animation without any explicit emotion label, style reference, or 3D capture of the target speaker. The discovery is a training recipe: a dual-branch student-teacher encoder pre-trained on 2D audio-visual clips, where masking is deliberately biased toward moments of high prosodic saliency measured by pitch deviation from neutral, learns speech features that predict when and how facial and head movements occur. The paper argues that visual supervision is indispensable, since rem

Load-bearing premise

The entire training and evaluation pipeline treats 3D meshes recovered by monocular reconstruction and then filtered and smoothed as ground truth; if those reconstructions drop or distort the subtle facial and head motions real humans produce, the reported improvements measure fidelity to the reconstruction pipeline rather than to true expressive motion.

Editorial extensions

If this is right

  • The pretrained encoder is a drop-in module: adding it to DiffPoseTalk and LSF-Animation improved their metrics, so other 3D talking-head frameworks can gain expressiveness without architectural redesign.
  • Near real-time (20.48 FPS on a single RTX 3090) expressive synthesis is reachable with audio-only input, with no emotion label, reference video, or style embedding required.
  • Emotion-aware masking organizes the learned representation by arousal (calm/sad/disgust versus happy/surprised/fearful), which the paper ties to smoother, less jittery animation.
  • Scaling pre-training data from 6 to 250 hours yields quickly diminishing returns, suggesting the audio-motion alignment prior is sample-efficient rather than data-hungry.
  • Jointly generating face and head as cascaded diffusion models, with the head generator conditioned on the face output, improves both lip synchronization and head-beat realism.

Reading between the lines

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

  • The reported gains are measured against pseudo-ground truth produced by monocular reconstruction, so the method's true ceiling is probably higher with high-fidelity 4D capture data; the hidden risk is that some reported advantages partially reflect which model best imitates SMIRK's smoothed outputs rather than human motion.
  • Because masking relies only on F0-derived saliency, emotions distinguished mainly by energy or spectral cues may be underrepresented; extending masking to additional prosodic signals is a natural testable upgrade.
  • The encoder outputs a motion-aligned representation rather than FLAME parameters, so it could plausibly transfer to 2D talking-head generation and avatar systems that do not use FLAME at all.
  • The arousal-organized latent structure suggests controllable emotion intensity or continuous emotion interpolation at inference time, an ability the paper does not demonstrate.
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 / 4 minor

Summary. The paper presents ETHead, a speech-driven 3D facial animation and head-movement generation method. It pre-trains a motion-aligned speech encoder via audio-visual self-distillation on large 2D talking-head datasets, using an F0-derived emotion saliency profile to bias masking toward emotionally salient segments. The resulting encoder is integrated into a diffusion-based FLAME generator through input feature modulation and output supervision in a contrastively learned speech-motion latent space. Experiments on RAVDESS, MEAD, and HDTF report improvements over DiffPoseTalk, LSF-Animation, and DEEPTalk on LVE, EVE, FDD, BA, and FID, plus a user study, and ablation results supporting the proposed components. The paper also claims the encoder is a transferable module for other 3D talking-head frameworks.

Significance. If the quantitative claims held, the paper would make a useful contribution: a method for transferring expressive motion priors from abundant 2D video to 3D animation, with a concrete masking mechanism and a transferable encoder. The ablations are thoughtfully designed and mostly consistent with the stated hypotheses; the t-SNE analysis and pretraining-scale study add value. However, the headline comparisons rest on pseudo-ground-truth reconstruction targets and single-run metrics, so the significance of the claimed 'substantial outperformance' is not yet established.

major comments (4)
  1. [Sec. IV-A, Eqs. (3)-(10), Sec. V] All training losses and all quantitative metrics (LVE, EVE, FDD, BA, FID) are computed against SMIRK-reconstructed, filtered, and smoothed meshes. The paper itself states in Sec. V that this pseudo-GT 'often underestimates subtle facial motions.' Consequently, Tables I and II may measure how well a model reproduces the reconstruction pipeline rather than genuine expressive human motion, especially because the same processed targets are used for training. Please add a held-out evaluation with real 3D ground truth or human-verified reconstructions, or otherwise show that the attenuation does not drive the reported gains.
  2. [Sec. IV-B, Tables I, II, V] The headline comparisons are single-run point estimates with no standard deviations or significance tests. Several differences are very small (e.g., Table I EVE: ours 1.266 vs. DiffPoseTalk* 1.276; Table II BA: ours 2.601 vs. DiffPoseTalk 2.592, a 0.35% difference). Table IV demonstrates that three-run statistics are feasible; please report mean±std and significance tests (or confidence intervals) for Tables I, II, and V before claiming substantial improvements.
  3. [Sec. IV-A (Evaluation Protocols)] The in-domain protocol is 'seen-subject, unseen-utterance': the first sentence of the three test speakers is included in the training set. The model therefore has access to the test speakers' facial structure and expressive style during training. This may inflate in-domain performance and weakens the claim of reproducing actor-specific emotional nuances as evidence of generalization. Please also report a fully unseen-subject split, or justify why the seen-subject protocol is the appropriate test for the paper's central claim.
  4. [Sec. IV-E (User Study)] The study uses 26 participants and reports preference percentages, but no per-criterion means, variances, confidence intervals, or significance tests. The statement that 'approximately 60% of participants rated our results as either indistinguishable from or even preferable to the tracked human motions' is not quantifiable without the underlying scale and distribution. Also, if the 'Ground Truth' is the same pseudo-GT rendering used in the quantitative evaluation, the study inherits the validity concern raised above. Please report full statistics and clarify the reference stimuli.
minor comments (4)
  1. [Appendix (Emotion-Aware Audio Masking)] The mixture weight between the saliency-based and uniform masking distributions is not specified. Please report this hyperparameter and the exact schedule for the visual masking ratio (0.1 to 0.6).
  2. [Table I] The caption says the better result between original and augmented variants is underlined, but underlining is not visible in the table; please use a clear marker.
  3. [Sec. III-C] The joint speech-motion latent space is motivated by [19], [48], [61]; since [61] is the authors' own earlier EcoFace work, please clarify the new contribution relative to that work.
  4. [Sec. IV-D, Fig. 7] The t-SNE analysis is qualitative; reporting emotion classification accuracy with confidence intervals would make the claim of improved separability quantitative.

Circularity Check

0 steps flagged · score 2.0 of 10

No significant circularity; minor self-citation and acknowledged pseudo-GT limitation do not make the derivation self-referential.

full rationale

The paper's central derivation chain is self-contained. The motion-aligned speech encoder is pre-trained via self-distillation on 2D audio-visual data (CelebV-HQ/CelebV-Text), then transferred as a frozen module to a 3D diffusion-based talking-head model trained on SMIRK pseudo-ground truth. The output-supervision loss Lemo (Eq. 11) is an internal regularizer and is explicitly not used as an evaluation metric; the reported metrics (LVE, EVE, FDD, BA, FID) are external geometry/statistics-based comparisons against the same pseudo-GT used for training, but they are not identical to training losses. The training losses (Eqs. 3-10) and evaluation metrics measure different quantities, so there is no 'Eq. X = Eq. Y by construction' reduction. The only self-citation is to EcoFace [61] for the general concept of joint speech-motion latent space, but the same concept is also credited to independent works [19], [48] and is implemented with contrastive learning (CPC loss [81]) rather than by directly reusing EcoFace's model. This self-citation is not load-bearing. The in-domain protocol deliberately includes test subjects' first sentences in training, and the paper transparently acknowledges this; it is a methodological choice, not circularity. The pseudo-GT limitation is explicitly admitted in Sec V ('the pseudo-GT from monocular reconstruction often underestimates subtle facial motions'), which tempers absolute expressiveness claims and applies equally to all compared methods; it is a validity concern, not a circular derivation. No fitted parameter is renamed as a prediction, and no uniqueness theorem or ansatz is smuggled in via self-citation. Thus the paper does not exhibit significant circularity.

Assumptions & free parameters 5 free parameters · 5 assumptions · 0 invented entities

The central empirical claims rest on a chain of domain assumptions about F0 as a proxy for emotional saliency, about pseudo-GT from SMIRK being reliable enough for training and evaluation, and about 2D video corpora containing speech-motion alignment. The main hand-set hyperparameters are the composite loss weights and masking schedules, all chosen on the benchmark domain without sensitivity analysis.

free parameters (5)
  • Composite loss weights = lambda_vtx_FM=2e6, lambda_vel_FM=1e7, lambda_acc_FM=1e5, lambda_pos_HM=0.5, lambda_vel_HM=50, lambda_acc_HM=5
    Hand-set in the appendix to balance reconstruction and kinematic losses; no sensitivity analysis is reported.
  • Audio masking ratio = 0.6
    Fixed masking ratio for the emotion-aware audio masking, chosen empirically.
  • Visual masking schedule = 0.1 to 0.6 (dynamic)
    Dynamic masking ratio for the visual branch, chosen empirically.
  • EMA momentum and fusion dropout = alpha=0.996, beta=0.3
    Hyperparameters for the self-distillation framework, affecting representation quality.
  • F0 saliency mixture weight = not reported
    The masking distribution mixes softmax-normalized saliency scores with a uniform distribution; the mixture weight is not specified, leaving an unreported tuning choice.
assumptions (5)
  • ad hoc to paper The SMIRK reconstruction, after filtering and smoothing, produces pseudo-ground-truth 3D meshes accurate enough to train and evaluate expressive animation.
    Load-bearing for all reported metrics; the paper itself concedes that pseudo-GT underestimates subtle facial motions (Sec V).
  • domain assumption Absolute standardized F0 deviation is a valid proxy for emotional saliency.
    Used to generate the emotion-aware masking distribution in Algorithm 1 and the Appendix.
  • domain assumption CelebV-HQ and CelebV-Text clips contain facial and head dynamics aligned with speech, so self-distillation transfers useful motion priors.
    The pretraining phase relies on this alignment being present in large-scale 2D talking videos.
  • domain assumption Stochastic modality dropout lets the retained speech encoder be used without visual input at inference.
    The architecture's inference-time behavior depends on the visual branch being replaceable by learnable placeholder tokens.
  • domain assumption WavLM and emotion2vec provide valid content and emotion representations that are complementary to the new encoder.
    The 3D generator uses these pretrained models as fixed input features, and the paper's ablations compare against them.

how reviews work

0 comments
Cite this review

Pith. "Pith review of ETHead: Generating Expressive 3D Facial Animation and Head Movement from Speech." pith.science (2026). https://pith.science/paper/MEKTQJXH

@misc{pith2026260801605,
  author       = {Pith},
  title        = {Pith review of: ETHead: Generating Expressive 3D Facial Animation and Head Movement from Speech},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/MEKTQJXH}},
  note         = {Machine review of arXiv:2608.01605}
}
read the original abstract

Generating expressive 3D talking heads solely from speech remains a significant challenge due to the scarcity of high-fidelity 3D data, which limits the modeling of complex emotional motion patterns. In this paper, we introduce \textbf{E}xpressive \textbf{T}alking \textbf{Head} (ETHead), a method for generating 3D facial and head motions that vividly align with the emotional content of input speech. To overcome the data limitations, we design a self-distillation framework that leverages large-scale 2D talking videos to pre-train a specialized speech encoder. By incorporating a novel emotion-modulated probabilistic masking mechanism, this framework aligns speech representations with expressive visual dynamics, allowing the encoder to extract features highly correlated with facial and head motions directly from audio. These features are then leveraged to guide 3D generation, enriching input cues and providing explicit supervision through a joint speech-motion latent space. Extensive experiments demonstrate that ETHead substantially outperforms state-of-the-art methods. Furthermore, our motion-aligned speech encoder can serve as a transferable module, offering a general solution for enhancing expressiveness in other 3D talking head animation frameworks. The project page is available at https://verdure-oss.github.io/ETHead.github.io/.

Figures

Figures reproduced from arXiv: 2608.01605 by the authors.

Figure 1
Figure 1. Given input speech, our method synthesizes emotion-coherent facial and head motions. The key is a motion-aligned speech encoder pre-trained [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Training framework of the motion-aligned speech encoder. Given a talking head video clip, the Student and Teacher networks separately extract modality-specific features followed by intra-network fusion. Random temporal masking is applied exclusively to the Student’s features, specifically targeting segments with intense emotional dynamics identified by prosody variations. The visual tokens are randomly dropped and r… view at source ↗
Figure 3
Figure 3. Overview of the proposed framework. (a) Generation Pipeline: The model extracts linguistic content and emotional features from input speech to condition two diffusion-based generators. In a cascaded manner, synthesized facial motion parameters are combined with speech features to drive the subsequent head motion generator. Training is supervised by reconstruction and kinematic consistency objectives. (b) Auxiliary M… view at source ↗
Figures from the paper (5 more)
Figure 4
Figure 4. Figure 4: Statistical analysis of head pose dynamics. Results are shown for a specific speaker in the 3D-RAVDESS test set. The boxes indicate the full range (minimum to maximum), with the center line denoting the median. Our method produces pitch, yaw, and their angular velocity…
Figure 5
Figure 5. Figure 5: Visualization of head pose dynamics. The left line charts display the temporal evolution of pitch and yaw angles alongside their corresponding angular velocities. Notable segments are highlighted with dashed boxes, represented by the rendered head poses on the right. C…
Figure 6
Figure 6. Figure 6: Visual comparison of facial motion during emotional speech. To facilitate comparison, all results are rendered in a canonical frontal view with zero head pose. The top five rows present in-domain results on 3D-RAVDESS, while the bottom four rows show out-of-domain resu…
Figure 7
Figure 7. Figure 7: t-SNE visualization of learned representations on the RAVDESS dataset. (a) Representations extracted by our motion-aligned speech encoder trained [PITH_FULL_IMAGE:figures/full_fig_p010_7.png]
Figure 8
Figure 8. Figure 8: User study results. Participants were asked to indicate their preferred result between our method and a competing approach under the specified evaluation criteria. The chart reports the corresponding preference percentages. gap between synthesized animations and real-w…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

92 extracted references · 74 canonical work pages

  1. [1]

    Emotalk: Speech-driven emotional disentanglement for 3d face animation,

    Z. Peng, H. Wu, Z. Song, H. Xu, X. Zhu, J. He, H. Liu, and Z. Fan, “Emotalk: Speech-driven emotional disentanglement for 3d face animation,” inProceedings of the IEEE/CVF International Conference on Computer Vision (ICCV). IEEE, 2023, pp. 20 630–20 640

  2. [2]

    Emotional speech-driven animation with content-emotion disentangle- ment,

    R. Danecek, K. Chhatre, S. Tripathi, Y . Wen, M. J. Black, and T. Bolkart, “Emotional speech-driven animation with content-emotion disentangle- ment,” inProceedings of the SIGGRAPH Asia. ACM, 2023, pp. 41:1– 41:13

  3. [3]

    DiffPoseTalk: Speech-driven stylistic 3d facial animation and head pose generation via diffusion models,

    Z. Sun, T. Lv, S. Ye, M. Lin, J. Sheng, Y .-H. Wen, M. Yu, and Y .-j. Liu, “DiffPoseTalk: Speech-driven stylistic 3d facial animation and head pose generation via diffusion models,”ACM Trans. Graph., vol. 43, no. 4, pp. 1–9, 2024

  4. [4]

    Adamesh: Personalized facial expressions and head poses for adaptive speech-driven 3d facial animation,

    L. Chen, W. Bao, S. Lei, B. Tang, Z. Wu, S. Kang, H. Huang, and H. Meng, “Adamesh: Personalized facial expressions and head poses for adaptive speech-driven 3d facial animation,”IEEE Trans. Multim., vol. 27, pp. 3598–3609, 2025

  5. [5]

    Imitator: Personalized speech-driven 3d facial animation,

    B. Thambiraja, I. Habibie, S. Aliakbarian, D. Cosker, C. Theobalt, and J. Thies, “Imitator: Personalized speech-driven 3d facial animation,” in Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV). IEEE, 2023, pp. 20 564–20 574

  6. [6]

    Omnihuman-1.5: Instilling an active mind in avatars via cognitive simulation,

    J. Jiang, W. Zeng, Z. Zheng, J. Yang, C. Liang, W. Liao, H. Liang, Y . Zhang, and M. Gao, “Omnihuman-1.5: Instilling an active mind in avatars via cognitive simulation,”arXiv preprint arXiv:2508.19209, 2025

  7. [7]

    Supervising 3d talking head avatars with analysis-by-audio-synthesis,

    R. Dan ˇeˇcek, C. Schmitt, S. Polikovsky, and M. J. Black, “Supervising 3d talking head avatars with analysis-by-audio-synthesis,”arXiv preprint arXiv:2504.13386, 2025

  8. [8]

    JALI: an animator- centric viseme model for expressive lip synchronization,

    P. Edwards, C. Landreth, E. Fiume, and K. Singh, “JALI: an animator- centric viseme model for expressive lip synchronization,”ACM Trans. Graph., vol. 35, no. 4, pp. 1–11, 2016

Show all 92 references
  1. [9]

    Massaro, M

    D. Massaro, M. Cohen, M. Tabain, J. Beskow, and R. Clark,Animated speech: research progress and applications. Cambridge University Press, 2012, p. 309–345

  2. [10]

    Dynamic units of visual speech,

    S. L. Taylor, M. Mahler, B.-J. Theobald, and I. Matthews, “Dynamic units of visual speech,” inProceedings of the ACM SIGGRAPH/Euro- graphics Symposium on Computer Animation, ser. SCA ’12. Goslar, DEU: Eurographics Association, 2012, p. 275–284

  3. [11]

    A practical and configurable lip sync method for games,

    Y . Xu, A. W. Feng, S. Marsella, and A. Shapiro, “A practical and configurable lip sync method for games,” inProceedings of Motion on Games, ser. MIG ’13. New York, NY , USA: Association for Computing Machinery, 2013, p. 131–140. [Online]. Available: https://doi.org/10.1145/25...

  4. [12]

    Speech-driven 3d face animation with composite and regional facial movements,

    H. Wu, S. Zhou, J. Jia, J. Xing, Q. Wen, and X. Wen, “Speech-driven 3d face animation with composite and regional facial movements,” in Proceedings of the 31st ACM International Conference on Multimedia. ACM, 2023, pp. 6822–6830

  5. [13]

    Ecavatar: 3d avatar facial animation with controllable identity and emotion,

    M. Yu, D. Pang, Z. Kang, Z. Sun, T. Lv, J. Sheng, R. Yi, Y . Wen, and Y . Liu, “Ecavatar: 3d avatar facial animation with controllable identity and emotion,” inProceedings of the 32nd ACM International Conference on Multimedia. ACM, 2024, pp. 10 468–10 476

  6. [14]

    S3: speech, script and scene driven head and eye animation,

    Y . Pan, R. Agrawal, and K. Singh, “S3: speech, script and scene driven head and eye animation,”ACM Trans. Graph., vol. 43, no. 4, pp. 47:1– 47:12, 2024

  7. [15]

    Expressive 3d facial animation generation based on local-to-global latent diffusion,

    W. Song, X. Wang, Y . Jiang, S. Li, A. Hao, X. Hou, and H. Qin, “Expressive 3d facial animation generation based on local-to-global latent diffusion,”IEEE Trans. Vis. Comput. Graph., vol. 30, no. 11, pp. 7397–7407, 2024

  8. [16]

    Learn2talk: 3d talking face learns from 2d talking face,

    Y . Zhuang, B. Cheng, Y . Cheng, Y . Jin, R. Liu, C. Li, X. Cheng, J. Liao, and J. Lin, “Learn2talk: 3d talking face learns from 2d talking face,” IEEE Trans. Vis. Comput. Graph., vol. 31, no. 9, pp. 5829–5841, 2025

  9. [17]

    Talkingstyle: Personalized speech-driven 3d facial animation with style preservation,

    W. Song, X. Wang, S. Zheng, S. Li, A. Hao, and X. Hou, “Talkingstyle: Personalized speech-driven 3d facial animation with style preservation,” IEEE Trans. Vis. Comput. Graph., vol. 31, no. 9, pp. 4682–4694, 2025

  10. [18]

    Emoface: Audio-driven emotional 3d face animation,

    C. Liu, Q. Lin, Z. Zeng, and Y . Pan, “Emoface: Audio-driven emotional 3d face animation,” inIEEE Conference Virtual Reality and 3D User Interfaces (VR). IEEE, 2024, pp. 387–397

  11. [19]

    Perceptually accurate 3d talking head generation: New definitions, speech-mesh representation, and evaluation metrics,

    L. Chae-Yeon, O. Hyun-Bin, H. EunGi, K. Sung-Bin, S. Nam, and T. Oh, “Perceptually accurate 3d talking head generation: New definitions, speech-mesh representation, and evaluation metrics,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CV...

  12. [20]

    Wav2sem: Plug-and- play audio semantic decoupling for 3d speech-driven facial animation,

    H. Li, J. Dai, X. Zhao, F. Zhou, J. Pan, and L. Li, “Wav2sem: Plug-and- play audio semantic decoupling for 3d speech-driven facial animation,” inIEEE/CVF Conference on Computer Vision and Pattern Recognition. Computer Vision Foundation / IEEE, 2025, pp. 183–192

  13. [21]

    Ot-talk: Animating 3d talking head with optimal transportation,

    X. Wang, X. Gao, X. Song, H. Yu, Z. Lin, L. Peng, and X. Gu, “Ot-talk: Animating 3d talking head with optimal transportation,” inProceedings of the 2025 International Conference on Multimedia Retrieval. ACM, 2025, pp. 1340–1349

  14. [22]

    Medtalk: Mul- timodal controlled 3d facial animation with dynamic emotions by disentangled embedding,

    C. Liu, Y . Pan, C. Ding, S. Rahardja, and X. Yang, “Medtalk: Mul- timodal controlled 3d facial animation with dynamic emotions by disentangled embedding,” inProceedings of the 33rd ACM International Conference on Multimedia. ACM, 2025, p. 7538–7547

  15. [23]

    Ptalker: Personal- ized speech-driven 3d talking head animation via style disentanglement and modality alignment,

    B. Wang, Y . Xu, H. Zhao, H. Zhang, and Z. Zhang, “Ptalker: Personal- ized speech-driven 3d talking head animation via style disentanglement and modality alignment,” inProceedings of the 33rd ACM International Conference on Multimedia. ACM, 2025, p. 10334–10342

  16. [24]

    Artalk: Speech- driven 3d head animation via autoregressive model,

    X. Chu, N. Goswami, Z. Cui, H. Wang, and T. Harada, “Artalk: Speech- driven 3d head animation via autoregressive model,” inProceedings of the SIGGRAPH Asia. ACM, 2025

  17. [25]

    Emovoca: Speech-driven emotional 3d talking heads,

    F. Nocentini, C. Ferrari, and S. Berretti, “Emovoca: Speech-driven emotional 3d talking heads,” inIEEE/CVF Winter Conference on Appli- cations of Computer Vision. IEEE, 2025, pp. 2859–2868

  18. [26]

    A semantic talking style space for speech-driven facial animation,

    Y . Chai, Y . Weng, T. Shao, and K. Zhou, “A semantic talking style space for speech-driven facial animation,”IEEE Trans. Vis. Comput. Graph., vol. 31, no. 12, pp. 10 801–10 814, 2025

  19. [27]

    Facexhubert: Text-less speech-driven e(x)pressive 3d facial animation synthesis using self-supervised speech representation learning,

    K. I. Haque and Z. Yumak, “Facexhubert: Text-less speech-driven e(x)pressive 3d facial animation synthesis using self-supervised speech representation learning,” inProceedings of the 25th International Con- ference on Multimodal Interaction (ICMI). ACM, 2023, pp. 282–291

  20. [28]

    Capture, learning, and synthesis of 3d speaking styles,

    D. Cudeiro, T. Bolkart, C. Laidlaw, A. Ranjan, and M. J. Black, “Capture, learning, and synthesis of 3d speaking styles,” inProceedings of the IEEE Computer Vision and Pattern Recognition (CVPR). IEEE, 2019, pp. 10 101–10 111

  21. [29]

    Meshtalk: 3d face animation from speech using cross-modality disen- tanglement,

    A. Richard, M. Zollh ¨ofer, Y . Wen, F. D. la Torre, and Y . Sheikh, “Meshtalk: 3d face animation from speech using cross-modality disen- tanglement,” inProceedings of the IEEE/CVF International Conference on Computer Vision (ICCV). IEEE, 2021, pp. 1153–1162

  22. [30]

    Unitalker: Scaling up audio- driven 3d facial animation through A unified model,

    X. Fan, J. Li, Z. Lin, W. Xiao, and L. Yang, “Unitalker: Scaling up audio- driven 3d facial animation through A unified model,” inProceedings of the European Conference on Computer Vision (ECCV), vol. 15099. Springer, 2024, pp. 204–221

  23. [31]

    Faceformer: Speech- driven 3d facial animation with transformers,

    Y . Fan, Z. Lin, J. Saito, W. Wang, and T. Komura, “Faceformer: Speech- driven 3d facial animation with transformers,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). IEEE, 2022, pp. 18 749–18 758

  24. [32]

    Selftalk: A self-supervised commutative training diagram to compre- hend 3d talking faces,

    Z. Peng, Y . Luo, Y . Shi, H. Xu, X. Zhu, H. Liu, J. He, and Z. Fan, “Selftalk: A self-supervised commutative training diagram to compre- hend 3d talking faces,” inProceedings of the 31st ACM International Conference on Multimedia (MM). ACM, 2023, pp. 5292–5301

  25. [33]

    Pmmtalk: Speech-driven 3d facial animation from complementary pseudo multi-modal features,

    T. Han, S. Gui, Y . Huang, B. Li, L. Liu, B. Zhou, N. Jiang, Q. Lu, R. Zhi, Y . Liang, D. Zhang, and J. Wan, “Pmmtalk: Speech-driven 3d facial animation from complementary pseudo multi-modal features,” IEEE Trans. Multim., vol. 27, pp. 2570–2581, 2025

  26. [34]

    Expclip: Bridging text and facial expressions via semantic alignment,

    Y . Zhong, H. Wei, P. Yang, and Z. Wang, “Expclip: Bridging text and facial expressions via semantic alignment,” inThirty-Eighth AAAI Conference on Artificial Intelligence (AAAI). AAAI Press, 2024, pp. 7614–7622

  27. [35]

    Mimic: Speaking style disentanglement for speech-driven 3d facial animation,

    H. Fu, Z. Wang, K. Gong, K. Wang, T. Chen, H. Li, H. Zeng, and W. Kang, “Mimic: Speaking style disentanglement for speech-driven 3d facial animation,” inThirty-Eighth AAAI Conference on Artificial Intelligence (AAAI). AAAI Press, 2024, pp. 1770–1777

  28. [36]

    Kmtalk: Speech-driven 3d facial animation with key motion embed- ding,

    Z. Xu, S. Gong, J. Tang, L. Liang, Y . Huang, H. Li, and S. Huang, “Kmtalk: Speech-driven 3d facial animation with key motion embed- ding,” in18th European Conference on Computer Vision (ECCV), vol. 15114. Springer, 2024, pp. 236–253

  29. [37]

    ScanTalk: 3d talking heads from unregistered scans,

    F. Nocentini, T. Besnier, C. Ferrari, S. Arguill `ere, S. Berretti, and M. Daoudi, “ScanTalk: 3d talking heads from unregistered scans,” in Proceedings of the 18th European Conference on Computer Vision (ECCV), vol. 15087. Springer, 2024, pp. 19–36

  30. [38]

    Deitalk: Speech-driven 3d facial animation with dynamic emotional intensity modeling,

    K. Shen, H. Xia, G. Geng, G. Geng, S. Xia, and Z. Ding, “Deitalk: Speech-driven 3d facial animation with dynamic emotional intensity modeling,” inProceedings of the 32nd ACM International Conference on Multimedia (MM). ACM, 2024, pp. 10 506–10 514

  31. [39]

    Audio-driven speech animation with text-guided expression,

    J. Noh, S. Junget al., “Audio-driven speech animation with text-guided expression,” inPacific Graphics 2024. AsiaGraphics Association, 2024

  32. [40]

    Memorytalker: Personalized speech- driven 3d facial animation via audio-guided stylization,

    H. K. Kim, S. Lee, and H. G. Kim, “Memorytalker: Personalized speech- driven 3d facial animation via audio-guided stylization,” inProceedings of the IEEE/CVF International Conference on Computer Vision (ICCV), 2025, pp. 11 241–11 251

  33. [41]

    Stylespeaker: Audio-enhanced fine- grained style modeling for speech-driven 3d facial animation,

    A. Yang, C. Liu, P. Xia, and J. Du, “Stylespeaker: Audio-enhanced fine- grained style modeling for speech-driven 3d facial animation,”arXiv preprint arXiv:2503.09852, 2025

  34. [42]

    Pestalk: Speech-driven 3d facial animation with personalized emotional styles,

    T. Han, B. Zhou, A. Liu, Y . Liang, D. Zhang, Z. Lei, and J. Wan, “Pestalk: Speech-driven 3d facial animation with personalized emotional styles,” inProceedings of the 33rd ACM International Conference on Multimedia (MM), 2025, pp. 7893–7901

  35. [43]

    Codetalker: Speech-driven 3d facial animation with discrete motion prior,

    J. Xing, M. Xia, Y . Zhang, X. Cun, J. Wang, and T. Wong, “Codetalker: Speech-driven 3d facial animation with discrete motion prior,” inPro- ceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). IEEE, 2023, pp. 12 780–12 790

  36. [44]

    Probabilistic speech-driven 3d facial motion synthesis: New bench- marks, methods, and applications,

    K. D. Yang, A. Ranjan, J. R. Chang, R. Vemulapalli, and O. Tuzel, “Probabilistic speech-driven 3d facial motion synthesis: New bench- marks, methods, and applications,” inIEEE/CVF Conference on Com- puter Vision and Pattern Recognition (CVPR). IEEE, 2024, pp. 27 284– 27 293

  37. [45]

    Multitalk: Enhancing 3d talking head generation across languages with multilingual video dataset,

    K. Sung-Bin, L. Chae-Yeon, G. Son, O. Hyun-Bin, J. Ju, S. Nam, and T. Oh, “Multitalk: Enhancing 3d talking head generation across languages with multilingual video dataset,” in25th Annual Conference of the International Speech Communication Association (Interspeech). ISCA, 2024

  38. [46]

    Probtalk3d: Non-deterministic emotion controllable speech-driven 3d facial animation synthesis using VQ-V AE,

    S. Wu, K. I. Haque, and Z. Yumak, “Probtalk3d: Non-deterministic emotion controllable speech-driven 3d facial animation synthesis using VQ-V AE,” inProceedings of the 17th ACM SIGGRAPH Conference on Motion, Interaction, and Games, (MIG). ACM, 2024, p. 15

  39. [47]

    Mmhead: Towards fine-grained multi-modal 3d facial animation,

    S. Wu, Y . Li, Y . Yan, H. Duan, Z. Liu, and G. Zhai, “Mmhead: Towards fine-grained multi-modal 3d facial animation,” inProceedings of the 32nd ACM International Conference on Multimedia (MM). ACM, 2024, pp. 7966–7975

  40. [48]

    Deeptalk: Dynamic emotion embedding for probabilistic speech-driven 3d face animation,

    J. Kim, J. Cho, J. Park, S. Hwang, D. E. Kim, G. Kim, and Y . Yu, “Deeptalk: Dynamic emotion embedding for probabilistic speech-driven 3d face animation,” inProceedings of the AAAI Conference on Artificial Intelligence (AAAI). AAAI Press, 2025, pp. 4275–4283

  41. [49]

    Prosodytalker: 3d visual speech animation via prosody decomposition,

    Z. Li, X. Lv, Q. Liu, Q. Meng, X. Sun, and S. Zhang, “Prosodytalker: 3d visual speech animation via prosody decomposition,” inProceedings of the AAAI Conference on Artificial Intelligence (AAAI). AAAI Press, 2025, pp. 5110–5118

  42. [50]

    Let’s chorus: Partner-aware hybrid song-driven 3d head animation,

    X. Xie, Z. Huang, W. Xu, P. Xiao, X. Xu, and H. Zhang, “Let’s chorus: Partner-aware hybrid song-driven 3d head animation,” inIEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). Computer Vision Foundation / IEEE, 2025, pp. 5467–5476

  43. [51]

    Facediffuser: Speech-driven 3d facial animation synthesis using diffusion,

    S. Stan, K. I. Haque, and Z. Yumak, “Facediffuser: Speech-driven 3d facial animation synthesis using diffusion,” inProceedings of the 16th ACM SIGGRAPH Conference on Motion, Interaction and Games (MIG). ACM, 2023, pp. 13:1–13:11. JOURNAL OF LATEX CLASS FILES, VOL. 14, NO. 8, A...

  44. [52]

    3diface: Synthesizing and editing holistic 3d facial animation,

    B. Thambiraja, M. Prinzler, S. Aliakbarian, D. Cosker, and J. Thies, “3diface: Synthesizing and editing holistic 3d facial animation,” in International Conference on 3D Vision (3DV). IEEE, 2025, pp. 1462– 1472

  45. [53]

    Diffspeaker: Speech-driven 3d facial animation with diffusion transformer,

    Z. Ma, X. Zhu, G. Qi, C. Qian, Z. Zhang, and Z. Lei, “Diffspeaker: Speech-driven 3d facial animation with diffusion transformer,”arXiv preprint arXiv:2402.05712, 2024

  46. [54]

    Diffusiontalker: Efficient and compact speech-driven 3d talking head via personalizer-guided distillation,

    P. Chen, X. Wei, M. Lu, H. Chen, and F. Tian, “Diffusiontalker: Efficient and compact speech-driven 3d talking head via personalizer-guided distillation,” inIEEE International Conference on Multimedia and Expo (ICME). IEEE, 2025, pp. 1–6

  47. [55]

    Facetalk: Audio-driven motion diffusion for neural parametric head models,

    S. Aneja, J. Thies, A. Dai, and M. Nießner, “Facetalk: Audio-driven motion diffusion for neural parametric head models,” inIEEE/CVF Con- ference on Computer Vision and Pattern Recognition (CVPR). IEEE, 2024, pp. 21 263–21 273

  48. [56]

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

    Q. Zhao, P. Long, Q. Zhang, D. Qin, H. Liang, L. Zhang, Y . Zhang, J. Yu, and L. Xu, “Media2face: Co-speech facial animation generation with multi-modality guidance,” inACM SIGGRAPH Conference Papers (SIGGRAPH). ACM, 2024, pp. 1–13

  49. [57]

    Glditalker: Speech-driven 3d facial animation with graph latent diffusion transformer,

    Y . Lin, Z. Fan, X. Wu, L. Xiong, X. Li, W. Kang, L. Peng, S. Lei, and H. Xu, “Glditalker: Speech-driven 3d facial animation with graph latent diffusion transformer,” inProceedings of the Thirty-Fourth International Joint Conference on Artificial Intelligence (IJCAI). ijcai.or...

  50. [58]

    Towards high-fidelity 3d talking avatar with personalized dynamic texture,

    X. Li, J. Wang, Y . Cheng, Y . Zeng, X. Ren, W. Zhu, W. Zhao, and Y . Yan, “Towards high-fidelity 3d talking avatar with personalized dynamic texture,” inIEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). Computer Vision Foundation / IEEE, 2025, pp. 204–214

  51. [59]

    Model see model do: Speech- driven facial animation with style control,

    Y . Pan, K. Singh, and L. G. Hafemann, “Model see model do: Speech- driven facial animation with style control,” inProceedings of the Special Interest Group on Computer Graphics and Interactive Techniques Con- ference Conference Papers (SIGGRAPH). ACM, 2025

  52. [60]

    V ASA-Rig: Audio-driven 3d facial animation with ’live’ mood dynamics in virtual reality,

    Y . Pan, C. Liu, S. Xu, S. Tan, and J. Yang, “V ASA-Rig: Audio-driven 3d facial animation with ’live’ mood dynamics in virtual reality,”IEEE Trans. Vis. Comput. Graph., vol. 31, no. 5, pp. 2416–2425, 2025

  53. [61]

    Ecoface: Audio-visual emotional co-disentanglement speech-driven 3d talking face generation,

    J. Xie, S. Zhang, M. Li, C. Lv, Z. Zhao, and F. Wu, “Ecoface: Audio-visual emotional co-disentanglement speech-driven 3d talking face generation,” inProceedings of the International Conference on Learning Representations (ICLR), 2025

  54. [62]

    Lsf- animation: Label-free speech-driven facial animation via implicit feature representation,

    X. Lu, C. Zhuang, C. Jin, Z. Lu, Y . Wang, W. Liu, and J. Xiao, “Lsf- animation: Label-free speech-driven facial animation via implicit feature representation,” inProceedings of the SIGGRAPH Asia 2025 Conference Papers, 2025, pp. 1–12

  55. [63]

    xada: Controllable and expressive audio-driven animation,

    S. Taylor, S. Medina, J. Windle, E. Alcusa S ´aez, and I. Matthews, “xada: Controllable and expressive audio-driven animation,” inProceedings of the SIGGRAPH Conference Papers (SIGGRAPH). ACM, 2025

  56. [64]

    wav2vec 2.0: A framework for self-supervised learning of speech representations,

    A. Baevski, Y . Zhou, A. Mohamed, and M. Auli, “wav2vec 2.0: A framework for self-supervised learning of speech representations,” in Proceedings of the Annual Conference on Neural Information Processing Systems (NeurIPS), 2020

  57. [65]

    Hubert: Self-supervised speech representation learning by masked prediction of hidden units,

    W. Hsu, B. Bolte, Y . H. Tsai, K. Lakhotia, R. Salakhutdinov, and A. Mohamed, “Hubert: Self-supervised speech representation learning by masked prediction of hidden units,”IEEE ACM Trans. Audio Speech Lang. Process., vol. 29, pp. 3451–3460, 2021

  58. [66]

    Masked autoencoders that listen,

    P. Huang, H. Xu, J. Li, A. Baevski, M. Auli, W. Galuba, F. Metze, and C. Feichtenhofer, “Masked autoencoders that listen,” inProceedings of the Annual Conference on Neural Information Processing Systems (NeurIPS), 2022

  59. [67]

    data2vec: A general framework for self-supervised learning in speech, vision and language,

    A. Baevski, W. Hsu, Q. Xu, A. Babu, J. Gu, and M. Auli, “data2vec: A general framework for self-supervised learning in speech, vision and language,” inProceedings of the International Conference on Machine Learning (ICML), vol. 162. PMLR, 2022, pp. 1298–1312

  60. [68]

    emotion2vec: Self-supervised pre-training for speech emotion rep- resentation,

    Z. Ma, Z. Zheng, J. Ye, J. Li, Z. Gao, S. Zhang, and X. Chen, “emotion2vec: Self-supervised pre-training for speech emotion rep- resentation,” inProceedings of the Findings of the Association for Computational Linguistics (ACL). ACL, 2024, pp. 15 747–15 760

  61. [69]

    Learning audio- visual speech representation by masked multimodal cluster prediction,

    B. Shi, W.-N. Hsu, K. Lakhotia, and A. Mohamed, “Learning audio- visual speech representation by masked multimodal cluster prediction,” inProceedings of the International Conference on Learning Represen- tations (ICLR), 2022

  62. [70]

    Self- supervised audio-visual speech representations learning by multimodal self-distillation,

    J.-X. Zhang, G. Wan, Z.-H. Ling, J. Pan, J. Gao, and C. Liu, “Self- supervised audio-visual speech representations learning by multimodal self-distillation,” inProceedings of the IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 2023, pp. 1–5

  63. [71]

    Head movements encode emotions during speech and song,

    S. R. Livingstone and C. Palmer, “Head movements encode emotions during speech and song,”Emotion, vol. 16, no. 3, p. 365, 2016

  64. [72]

    The effects of visual beats on prosodic prominence: Acoustic analyses, auditory perception and visual percep- tion,

    E. Krahmer and M. Swerts, “The effects of visual beats on prosodic prominence: Acoustic analyses, auditory perception and visual percep- tion,”Journal of memory and language, vol. 57, no. 3, pp. 396–414, 2007

  65. [73]

    Emerging properties in self-supervised vision transformers,

    M. Caron, H. Touvron, I. Misra, H. J ´egou, J. Mairal, P. Bojanowski, and A. Joulin, “Emerging properties in self-supervised vision transformers,” inProceedings of the IEEE/CVF International Conference on Computer Vision (ICCV). IEEE, 2021, pp. 9630–9640

  66. [74]

    Image BERT pre-training with online tokenizer,

    J. Zhou, C. Wei, H. Wang, W. Shen, C. Xie, A. L. Yuille, and T. Kong, “Image BERT pre-training with online tokenizer,” inProceedings of the International Conference on Learning Representations (ICLR). Open- Review.net, 2022

  67. [75]

    Learning a model of facial shape and expression from 4D scans,

    T. Li, T. Bolkart, M. J. Black, H. Li, and J. Romero, “Learning a model of facial shape and expression from 4D scans,”ACM Trans. Graph., vol. 36, no. 6, pp. 194:1–194:17, 2017

  68. [76]

    On the continuity of rotation representations in neural networks,

    Y . Zhou, C. Barnes, J. Lu, J. Yang, and H. Li, “On the continuity of rotation representations in neural networks,” inProceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR). Computer Vision Foundation / IEEE, 2019, pp. 5745–5753

  69. [77]

    Wavlm: Large-scale self-supervised pre- training for full stack speech processing,

    S. Chen, C. Wang, Z. Chen, Y . Wu, S. Liu, Z. Chen, J. Li, N. Kanda, T. Yoshioka, X. Xiao, J. Wu, L. Zhou, S. Ren, Y . Qian, Y . Qian, J. Wu, M. Zeng, X. Yu, and F. Wei, “Wavlm: Large-scale self-supervised pre- training for full stack speech processing,”IEEE J. Sel. Top. Signa...

  70. [78]

    Denoising diffusion probabilistic mod- els,

    J. Ho, A. Jain, and P. Abbeel, “Denoising diffusion probabilistic mod- els,” inProceedings of the 34th International Conference on Neural Information Processing Systems (NeurIPS), 2020, pp. 6840–6851

  71. [79]

    Arbitrary style transfer in real-time with adaptive instance normalization,

    X. Huang and S. Belongie, “Arbitrary style transfer in real-time with adaptive instance normalization,” inProceedings of the IEEE Interna- tional Conference on Computer Vision (ICCV). IEEE Computer Society, 2017, pp. 1510–1519

  72. [80]

    6d rotation representation for unconstrained head pose estimation,

    T. Hempel, A. A. Abdelrahman, and A. Al-Hamadi, “6d rotation representation for unconstrained head pose estimation,” inProceedings of the IEEE International Conference on Image Processing (ICIP). IEEE, 2022, pp. 2496–2500

  73. [81]

    Representation learning with contrastive predictive coding,

    A. v. d. Oord, Y . Li, and O. Vinyals, “Representation learning with contrastive predictive coding,”arXiv preprint arXiv:1807.03748, 2018

  74. [82]

    Celebv-hq: A large-scale video facial attributes dataset,

    H. Zhu, W. Wu, W. Zhu, L. Jiang, S. Tang, L. Zhang, Z. Liu, and C. C. Loy, “Celebv-hq: A large-scale video facial attributes dataset,” in Proceedings of the European Conference on Computer Vision (ECCV), vol. 13667. Springer, 2022, pp. 650–667

  75. [83]

    Celebv- text: A large-scale facial text-video dataset,

    J. Yu, H. Zhu, L. Jiang, C. C. Loy, W. Cai, and W. Wu, “Celebv- text: A large-scale facial text-video dataset,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). IEEE, 2023, pp. 14 805–14 814

  76. [84]

    3d facial expressions through analysis- by-neural-synthesis,

    G. Retsinas, P. P. Filntisis, R. Danecek, V . F. Abrevaya, A. Roussos, T. Bolkart, and P. Maragos, “3d facial expressions through analysis- by-neural-synthesis,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). IEEE, 2024, pp. 2490–2501

  77. [85]

    The ryerson audio-visual database of emotional speech and song (ravdess): A dynamic, multimodal set of facial and vocal expressions in north american english,

    S. R. Livingstone and F. A. Russo, “The ryerson audio-visual database of emotional speech and song (ravdess): A dynamic, multimodal set of facial and vocal expressions in north american english,”PloS one, vol. 13, no. 5, p. e0196391, 2018

  78. [86]

    MEAD: A large-scale audio-visual dataset for emotional talking-face generation,

    K. Wang, Q. Wu, L. Song, Z. Yang, W. Wu, C. Qian, R. He, Y . Qiao, and C. C. Loy, “MEAD: A large-scale audio-visual dataset for emotional talking-face generation,” inProceedings of the European Conference on Computer Vision (ECCV), vol. 12366. Springer, 2020, pp. 700–717

  79. [87]

    Flow-guided one-shot talking face generation with a high-resolution audio-visual dataset,

    Z. Zhang, L. Li, Y . Ding, and C. Fan, “Flow-guided one-shot talking face generation with a high-resolution audio-visual dataset,” inProceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR). Computer Vision Foundation / IEEE, 2021, pp. 3661–3670

  80. [88]

    Bailando: 3d dance generation by actor-critic GPT with choreographic memory,

    L. Siyao, W. Yu, T. Gu, C. Lin, Q. Wang, C. Qian, C. C. Loy, and Z. Liu, “Bailando: 3d dance generation by actor-critic GPT with choreographic memory,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). IEEE, 2022, pp. 11 040– 11 049

  81. [89]

    A technique for the measurement of attitudes

    R. Likert, “A technique for the measurement of attitudes.”Archives of psychology, 1932

  82. [90]

    Autoregressive image generation without vector quantization,

    T. Li, Y . Tian, H. Li, M. Deng, and K. He, “Autoregressive image generation without vector quantization,” inProceedings of the Annual Conference on Neural Information Processing Systems (NeurIPS), 2024

  83. [91]

    Flow matching for generative modeling,

    Y . Lipman, R. T. Q. Chen, H. Ben-Hamu, M. Nickel, and M. Le, “Flow matching for generative modeling,” inProceedings of the International Conference on Learning Representations (ICLR). OpenReview.net, 2023. JOURNAL OF LATEX CLASS FILES, VOL. 14, NO. 8, AUGUST 2021 15

  84. [92]

    Fmdistance: A fast and effective distance function for motion capture data,

    K. Onuma, C. Faloutsos, and J. K. Hodgins, “Fmdistance: A fast and effective distance function for motion capture data,” inProceedings of the Annual Conference of the European Association for Computer Graphics (Eurographics). Eurographics Association, 2008, pp. 83–86. APPENDIX...

Pith tools

Reviewed August 5, 2026 · model on record in the stance chip above.