Pith. sign in

REVIEW 3 major objections 6 minor 3 cited by

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

T0 review · 3 major / 6 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read This paper claims that identity leakage and rendering artifacts in GAN-based talking head generation share one cause and one fix: the 4-th layer feature of the driven image carries identity, and reusing that feature both removes leakage…

desk verdict A competent, incremental talking-head paper whose f^d_4 repurposing idea is genuinely neat, but the load-bearing layer claim is qualitatively supported and the code is missing; send it to review, require the supplementary evidence in the main text. read the letter →

arxiv 2507.01390 v1 pith:J75CI4BG submitted 2025-07-02 cs.CV

Shuai Tan , Bill Gong , Bin Ji , Ye Pan This is my paper
classification cs.CV
keywords talkingheadgenerationidentityleakagerenderingartifactsGAN-basedanimationmotiondisentanglementmemorynetworkfacialexpressionreal-timesynthesis
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 tries to establish that the two worst failure modes of GAN-based talking head generation, identity leakage and rendering artifacts, can be solved together without losing real-time speed or decoupled control. Its key empirical finding is that the identity information responsible for leakage sits in the 4-th layer feature of the driven image, and that the same leaked information can supplement missing details when rendering extreme poses and expressions. FixTalk builds on this with two lightweight modules: EMI strips identity from motion features, and EDI stores and retrieves the identity-bearing features to repair artifacts. If correct, the result is a real-time GAN-based talking head generator that preserves the source identity and renders extreme expressions faithfully, whereas previous GAN methods typically trade one of these for the other.

What carries the argument

The load-bearing object is the driven image's 4-th layer encoder feature $f^d_4$, which the paper identifies as the carrier of identity leakage. FixTalk's two modules act on it: EMI uses a Q-Former-style learnable query in cross-attention to pull motion-specific content from $f^d_4$ and combine multi-scale features, with a disentanglement loss on source versus driven motion features; EDI compresses $f^d_4$ into tokens, stores them in a dual memory network aligned by KL divergence, and retrieves the identity-matched token at inference to supply missing details through multi-head cross-attention. The claim is that these two lightweight, plug-and-play modules let a GAN-based model retain real-time speed and decoupled control while eliminating leakage and artifacts.

What would settle it

Replace $f^d_4$ with the corresponding source feature in FixTalk and measure both identity preservation (CSIM against the driven identity) and motion fidelity; if leakage persists or motion accuracy collapses, the concentration claim fails. The paper's own conclusion depends on supplementary layer-wise ablations, so running those ablations on multiple baselines and identities is the direct check.

Watch

Extended reading notes

Core claim

On its own terms, the paper's discovery is that identity leakage in GAN-based talking head generation is not a byproduct to be suppressed but a resource to be redirected. Through systematic replacement of intermediate variables in the baseline EDTalk model, the paper claims to identify the 4-th layer feature $f^d_4$ of the driven image as the single variable that carries the identity information causing leakage; removing it prevents the driven identity from bleeding into the source appearance, while storing and reusing it in a self-driven context restores details lost in extreme poses and exaggerated expressions. FixTalk operationalizes this by isolating motion-specific content from $f^d_4$ in EMI and by using a leakage memory network in EDI that retrieves identity-matched detail features at inference. The paper argues this yields superior quantitative and qualitative performance against recent GAN-based and diffusion-based methods, while maintaining real-time inference and decoupled control.

Load-bearing premise

Everything rests on the empirical claim that identity leakage is concentrated in the single 4-th layer feature $f^d_4$ of the driven image and that removing or repurposing this feature does not damage motion fidelity; this is supported by qualitative replacement experiments and deferred ablations in the supplementary material, not by a proof, and may not transfer to other model architectures or identities.

Editorial extensions

If this is right

  • FixTalk simultaneously mitigates identity leakage and rendering artifacts while preserving real-time inference and decoupled control, giving GAN-based talking heads a path to image quality previously seen only in slower diffusion methods.
  • EMI's disentanglement loss on source versus driven motion features provides a concrete training signal that prevents driven identity from contaminating the motion representation.
  • EDI's memory network lets a cross-driven generator recall identity-matched detail features, so extreme poses and exaggerated expressions can be rendered without copying the driven identity into the source appearance.
  • Because both modules are plug-and-play, other GAN-based animators such as LIA can inherit the same fix, as the paper demonstrates in a comparative experiment.
  • In audio-driven settings, FixTalk retains high emotion accuracy and lip synchronization while improving identity preservation and reducing artifacts relative to its baseline.

Reading between the lines

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

  • If $f^d_4$ is indeed the main leakage carrier, the same two-module recipe could be transplanted to other GAN-based talking heads with minimal retraining; the paper's LIA experiment is suggestive but only tests one additional baseline.
  • The memory-based detail retrieval implies a dependence on the identity distribution seen during training: for a truly novel identity outside that distribution, the retrieved feature may be a poor match, which would be a testable failure mode.
  • The paper's replacement-variable diagnostic, swapping intermediate features between source and driven images, could become a standard way to localize identity leakage in future face animation models.
  • Since FixTalk still cannot infer natural emotions or subtle micro-expressions directly from audio, a natural next step would be combining this leakage-taming mechanism with a learned emotion prior, though the paper does not explore that direction.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 6 minor

Summary. The paper presents FixTalk, a GAN-based talking head generation framework built on EDTalk, aimed at simultaneously mitigating identity leakage and rendering artifacts while preserving real-time inference and decoupled control. The authors argue that identity leakage stems from identity information in the driven image's intermediate feature f^d_4, and that this same feature can be repurposed for detail recovery. FixTalk comprises EMI, which uses a learnable-query cross-attention extractor to distill motion features from f^d_4 and other multi-scale features under a disentanglement loss, and EDI, which stores compressed f^d_4 tokens in a dual memory network during training and retrieves them at inference to supplement rendering details. The method is evaluated on MEAD and HDTF for video- and audio-driven settings, with quantitative metrics, a user study, and an efficiency comparison showing 27.6 FPS on A100.

Significance. The paper's central contribution is a concrete, feature-level diagnosis of identity leakage in GAN-based talking heads and a mechanism that converts leaked identity information into a detail-recovery signal. If the identification of f^d_4 as the necessary carrier is correct, FixTalk offers a lightweight, plug-and-play remedy that preserves the efficiency and disentangled control of the underlying GAN baseline, which is a nontrivial advantage over diffusion-based alternatives. The authors provide extensive comparisons, a user study, efficiency analysis, and qualitative demonstrations on extreme poses and expressions, and they use publicly available training data. The main weakness is that the load-bearing premise about f^d_4 is not quantitatively demonstrated in the main text, and several quantitative claims lack uncertainty estimates.

major comments (3)
  1. [Sec. 3.2, Fig. 3] The claim that f^d_4 alone is 'necessary' to convey identity leakage is established only qualitatively in the main text; the layer-by-layer replacement study is deferred to the supplementary. Provide a quantitative ablation (e.g., CSIM and artifact metrics for each replaced layer) showing that replacing f^d_4 reproduces the effects of replacing f^d_{1:5}, and that no other layer carries comparable identity information. Without this, the design of both EMI and EDI rests on an unsupported premise.
  2. [Sec. 4.2, Eq. (9), and Sec. 5.2] EDI retrieves details from a memory trained on the training set, but HDTF evaluation uses source identities not seen during training. The paper should justify or empirically verify that the retrieval generalizes across identities, e.g., by analyzing the nearest retrieved slots for unseen identities or ablating memory retrieval on held-out identities.
  3. [Tables 1 and 3] All quantitative results are point estimates without standard errors or significance tests. Given the small margins (e.g., PSNR 27.164 vs. 27.054 for LivePor; SyncConf 8.009 vs. 8.115 for EDTalk), the claimed superiority is not statistically supported. Report variance across seeds or evaluation subsets, or provide significance tests.
minor comments (6)
  1. [Sec. 5.1] 'CSIM for accessing identity leakage' should be 'assessing'.
  2. [Sec. 4.2] 'decompressor decompressor Λ' contains a duplicated word.
  3. [Figure 4 caption] 'Multi-Head Cross-Attention mechanis' is truncated; complete the word 'mechanism'.
  4. [Eqs. (5) and (7)] The norm notation is ambiguous; e.g., '∥f dπ ∥2 · mi_d / ∥...∥' should be written with explicit division.
  5. [Table 3] The SyncConf of FixTalk (8.009) is slightly lower than EDTalk (8.115); the sentence 'best performance across most metrics' should acknowledge this exception.
  6. [Sec. 5.2] The symbols 'X-For#' and 'FYE#' are inconsistent; use a consistent marker convention.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: FixTalk's losses and memory mechanism do not reduce by construction to its evaluation metrics, and the self-cited baseline is a design choice rather than load-bearing evidence.

full rationale

The derivation chain is not circular at the level of the paper's own equations. The load-bearing premise that f^d_4 carries identity leakage is an empirical observation from the replacement experiments in Sec. 3.2, not a definition of identity leakage or of f^d_4. The EMI disentanglement loss directly minimizes cos(z_s, z_d) between motion features, but the reported identity metric CSIM measures cosine similarity between face-recognition embeddings of the source and the generated image; the two quantities are distinct, so the identity result is not forced by the loss term. Similarly, EDI's memory losses (L_d-mem, L_align) train a retrieval mechanism, but the rendering metrics (NIQE, CPBD, PSNR, FID) and the user study are not identical to those losses, so the artifact improvements are not by construction. The paper evaluates against external benchmarks (HDTF and a held-out MEAD partition) and against multiple independent baselines, including diffusion-based methods. The choice of EDTalk, the authors' own prior work, as the base model is a design decision and the self-citation is not used to justify the central claim; the comparison includes many non-overlapping methods. The main weakness is generalizability of the f^d_4 premise and the memory to unseen identities, which is a correctness risk, not circularity.

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

The central claims rest on empirical observations about the authors' own baseline, unproved assumptions about the separability of identity and motion, and a memory retrieval assumption. No new physical entities are introduced; the free parameters are hyperparameters required by the proposed losses and memory.

free parameters (4)
  • disentanglement margin xi = not reported
    In Eq. (4), L_dis = max(0, cos(z_s, z_d) - xi); xi controls the strictness of identity decoupling and is set by hand. No value is given in the main text.
  • number of memory slots S = not reported
    The EDI memory networks M_d and M_m-s each contain S slots; S is chosen by hand and not reported in the main text.
  • EMI extractor depth N = not reported
    The lightweight extractor P uses N stacked cross-attention and FFN layers; depth N is a hyperparameter and not reported in the main text.
  • loss weights for L_dis, L_d-mem, L_align, L_rec, L_per, L_adv = not reported
    The final objective combines multiple losses, but their balancing weights are not specified in the main text, so the training objective has unreported hyperparameters.
assumptions (4)
  • ad hoc to paper The 4-th layer feature f^d_4 of the driven image is the main carrier of identity leakage and the necessary and sufficient feature for detail supplementation.
    Sec. 3.2 and Fig. 3(d)(e) support this with qualitative examples; the claim is an empirical finding on the authors' own baseline and is the design basis for both EMI and EDI.
  • domain assumption Minimizing cosine similarity between source and driven motion features removes identity information without removing useful motion information.
    Eq. (4) imposes this directly; the paper offers no proof that identity and motion live in orthogonal subspaces, and in the self-driven setting source and driven share identity.
  • ad hoc to paper A memory network trained on self-driven pairs can retrieve identity-appropriate details for cross-driven inference via KL-aligned addresses.
    EDI assumes f^d_pi can be inferred from {z_d_s, f^s_pi} and that memory slots generalize to unseen identities; no evidence beyond the reported metrics.
  • domain assumption Rendering artifacts in extreme poses are caused primarily by missing details that can be added from identity-leaked features.
    Sec. 3.2 states this as an interpretation of the self-driven versus cross-driven gap; it motivates EDI but is not independently established.

how reviews work

0 comments
Cite this review

Pith. "Pith review of FixTalk: Taming Identity Leakage for High-Quality Talking Head Generation in Extreme Cases." pith.science (2026). https://pith.science/paper/J75CI4BG

@misc{pith2026250701390,
  author       = {Pith},
  title        = {Pith review of: FixTalk: Taming Identity Leakage for High-Quality Talking Head Generation in Extreme Cases},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/J75CI4BG}},
  note         = {Machine review of arXiv:2507.01390}
}
read the original abstract

Talking head generation is gaining significant importance across various domains, with a growing demand for high-quality rendering. However, existing methods often suffer from identity leakage (IL) and rendering artifacts (RA), particularly in extreme cases. Through an in-depth analysis of previous approaches, we identify two key insights: (1) IL arises from identity information embedded within motion features, and (2) this identity information can be leveraged to address RA. Building on these findings, this paper introduces FixTalk, a novel framework designed to simultaneously resolve both issues for high-quality talking head generation. Firstly, we propose an Enhanced Motion Indicator (EMI) to effectively decouple identity information from motion features, mitigating the impact of IL on generated talking heads. To address RA, we introduce an Enhanced Detail Indicator (EDI), which utilizes the leaked identity information to supplement missing details, thus fixing the artifacts. Extensive experiments demonstrate that FixTalk effectively mitigates IL and RA, achieving superior performance compared to state-of-the-art methods.

Figures

Figures reproduced from arXiv: 2507.01390 by the authors.

Figure 1
Figure 1. Example animations generated by FixTalk. Given an identity source, FixTalk supports decoupled control over lip sync, head pose, and emotional expressions from their respective sources. As presented in Tab. 4, FixTalk achieves real-time generation. In particular, FixTalk demonstrates superior performance with high-quality rendering in extreme poses and expressive expressions. Abstract Talking head generation is gaini… view at source ↗
Figure 2
Figure 2. Illustration of identity leakage and rendering artifact. [PITH_FULL_IMAGE:figures/full_fig_p002_2.png] view at source ↗
Figure 3
Figure 3. Latent feature investigation of Baseline. (b) presents the overview of Baseline. The source image [PITH_FULL_IMAGE:figures/full_fig_p003_3.png] view at source ↗
Figures from the paper (4 more)
Figure 4
Figure 4. Figure 4: The illustration of FixTalk. (a) FixTalk is built upon Baseline (covered by a gray mask). (b) Unlike Baseline, which directly [PITH_FULL_IMAGE:figures/full_fig_p005_4.png]
Figure 5
Figure 5. Figure 5: Qualitative comparisons in video-driven setting. See full comparison in supplementary material ( [PITH_FULL_IMAGE:figures/full_fig_p007_5.png]
Figure 6
Figure 6. Figure 6: Qualitative comparisons in audio-driven setting. See full comparison in supplementary material ( [PITH_FULL_IMAGE:figures/full_fig_p008_6.png]
Figure 8
Figure 8. Figure 8: Comparison between LIA wo/w FixTalk. LIA [71] as baseline. As presented in [PITH_FULL_IMAGE:figures/full_fig_p008_8.png]

Discussion (0). Sign in to comment.

Forward citations

Cited by 3 Pith papers

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

  1. Archon: A Unified Multimodal Model for Holistic Digital Human Generation

    cs.CV 2026-05 unverdicted novelty 5.0 of 10

    Archon unifies seven modalities via modality-specific tokenizers and an autoregressive backbone pretrained on 72 tasks, plus a 4x-efficient video reparameterization and stepwise 'Thinking in Modality' procedure, and r...

  2. PortraitDirector: A Hierarchical Disentanglement Framework for Controllable and Real-time Facial Reenactment

    cs.CV 2026-04 unverdicted novelty 5.0 of 10

    PortraitDirector uses hierarchical disentanglement of spatial physical motions and semantic emotions to deliver controllable, high-fidelity real-time facial reenactment at 20 FPS.

  3. EDTalk++: Full Disentanglement for Controllable Talking Head Synthesis

    cs.CV 2025-08 conditional novelty 5.0 of 10

    EDTalk++ disentangles talking-head video into four orthogonal motion banks (mouth, pose, eyes, expression) and drives them from either video or audio inputs.

Reference graph

Works this paper leans on

100 extracted references · 46 canonical work pages · cited by 3 Pith papers

  1. [1]

    Magicdance: Realistic human dance video gen- eration with motions & facial expressions transfer

    Di Chang, Yichun Shi, Quankai Gao, Jessica Fu, Hongyi Xu, Guoxian Song, Qing Yan, Xiao Yang, and Mohammad So- leymani. Magicdance: Realistic human dance video gen- eration with motions & facial expressions transfer. arXiv preprint arXiv:2311.12052, 2023. 3

    arXiv 2023

  2. [2]

    Anifacediff: High-fidelity face reenactment via fa- cial parametric conditioned diffusion models

    Ken Chen, Sachith Seneviratne, Wei Wang, Dongting Hu, Sanjay Saha, Md Tarek Hasan, Sanka Rasnayaka, Tamasha Malepathirana, Mingming Gong, and Saman Hal- gamuge. Anifacediff: High-fidelity face reenactment via fa- cial parametric conditioned diffusion models. arXiv preprint arXiv:2406.13272, 2024. 3

    arXiv 2024

  3. [3]

    Lip movements generation at a glance

    Lele Chen, Zhiheng Li, Ross K Maddox, Zhiyao Duan, and Chenliang Xu. Lip movements generation at a glance. In Proceedings of the European Conference on Computer Vi- sion (ECCV), pages 520–535, 2018. 3

    2018

  4. [4]

    Hierarchical cross-modal talking face generation with dynamic pixel-wise loss

    Lele Chen, Ross K Maddox, Zhiyao Duan, and Chenliang Xu. Hierarchical cross-modal talking face generation with dynamic pixel-wise loss. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages 7832–7841, 2019. 1, 3, 6

    2019

  5. [5]

    Talking-head generation with rhyth- mic head motion

    Lele Chen, Guofeng Cui, Celong Liu, Zhong Li, Ziyi Kou, Yi Xu, and Chenliang Xu. Talking-head generation with rhyth- mic head motion. In European Conference on Computer Vi- sion, pages 35–51. Springer, 2020. 3

    2020

  6. [6]

    Vast: Vivify your talk- ing avatar via zero-shot expressive facial style transfer

    Liyang Chen, Zhiyong Wu, Runnan Li, Weihong Bao, Jun Ling, Xu Tan, and Sheng Zhao. Vast: Vivify your talk- ing avatar via zero-shot expressive facial style transfer. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 2977–2987, 2023. 3

    2023

  7. [7]

    Echomimic: Lifelike audio-driven por- trait animations through editable landmark conditions

    Zhiyuan Chen, Jiajiong Cao, Zhiquan Chen, Yuming Li, and Chenguang Ma. Echomimic: Lifelike audio-driven por- trait animations through editable landmark conditions. arXiv preprint arXiv:2407.08136, 2024. 3, 7, 8

    arXiv 2024

  8. [8]

    Out of time: auto- mated lip sync in the wild

    Joon Son Chung and Andrew Zisserman. Out of time: auto- mated lip sync in the wild. In Computer Vision–ACCV 2016 Workshops: ACCV 2016 International Workshops, Taipei, Taiwan, November 20-24, 2016, Revised Selected Papers, Part II 13, pages 251–263. Springer, 2017. 6

    2016

Show all 100 references
  1. [9]

    Hallo2: Long-duration and high-resolution audio-driven portrait im- age animation

    Jiahao Cui, Hui Li, Yao Yao, Hao Zhu, Hanlin Shang, Kaihui Cheng, Hang Zhou, Siyu Zhu, and Jingdong Wang. Hallo2: Long-duration and high-resolution audio-driven portrait im- age animation. arXiv preprint arXiv:2410.07718, 2024. 3

    arXiv

  2. [10]

    Hallo3: Highly dynamic and realistic portrait image an- imation with diffusion transformer networks

    Jiahao Cui, Hui Li, Yun Zhan, Hanlin Shang, Kaihui Cheng, Yuqi Ma, Shan Mu, Hang Zhou, Jingdong Wang, and Siyu Zhu. Hallo3: Highly dynamic and realistic portrait image an- imation with diffusion transformer networks. arXiv preprint arXiv:2412.00733, 2024. 3, 8

    2024 arXiv

  3. [11]

    Speech-driven facial animation using cas- caded gans for learning of motion and texture

    Dipanjan Das, Sandika Biswas, Sanjana Sinha, and Brojesh- war Bhowmick. Speech-driven facial animation using cas- caded gans for learning of motion and texture. In Computer Vision–ECCV 2020: 16th European Conference, Glasgow, UK, August 23–28, 2020, Proceedings, Part XXX 16 , ...

    2020

  4. [12]

    Megaportraits: One-shot megapixel neural head avatars

    Nikita Drobyshev, Jenya Chelishev, Taras Khakhulin, Alek- sei Ivakhnenko, Victor Lempitsky, and Egor Zakharov. Megaportraits: One-shot megapixel neural head avatars. In Proceedings of the 30th ACM International Conference on Multimedia, pages 2663–2671, 2022. 2

    2022

  5. [13]

    Emoportraits: Emotion-enhanced multimodal one-shot head avatars

    Nikita Drobyshev, Antoni Bigata Casademunt, Konstantinos V ougioukas, Zoe Landgraf, Stavros Petridis, and Maja Pan- tic. Emoportraits: Emotion-enhanced multimodal one-shot head avatars. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , page...

    2024

  6. [14]

    Efficient emotional adaptation for audio-driven talking-head generation

    Yuan Gan, Zongxin Yang, Xihang Yue, Lingyun Sun, and Yi Yang. Efficient emotional adaptation for audio-driven talking-head generation. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 22634– 22645, 2023. 7

    2023

  7. [15]

    Generative adversarial networks

    Ian Goodfellow, Jean Pouget-Abadie, Mehdi Mirza, Bing Xu, David Warde-Farley, Sherjil Ozair, Aaron Courville, and Yoshua Bengio. Generative adversarial networks. Commu- nications of the ACM, 63(11):139–144, 2020. 2

    2020

  8. [16]

    Stylesync: High-fidelity generalized and personalized lip sync in style-based genera- tor

    Jiazhi Guan, Zhanwang Zhang, Hang Zhou, Tianshu Hu, Kaisiyuan Wang, Dongliang He, Haocheng Feng, Jingtuo Liu, Errui Ding, Ziwei Liu, et al. Stylesync: High-fidelity generalized and personalized lip sync in style-based genera- tor. In Proceedings of the IEEE/CVF Conference on C...

  9. [17]

    Resyncer: Rewiring style-based generator for unified audio-visually synced fa- cial performer

    Jiazhi Guan, Zhiliang Xu, Hang Zhou, Kaisiyuan Wang, Shengyi He, Zhanwang Zhang, Borong Liang, Haocheng Feng, Errui Ding, Jingtuo Liu, et al. Resyncer: Rewiring style-based generator for unified audio-visually synced fa- cial performer. arXiv preprint arXiv:2408.03284, 2024. 3

    2024 arXiv

  10. [18]

    Livepor- trait: Efficient portrait animation with stitching and retarget- ing control

    Jianzhu Guo, Dingyun Zhang, Xiaoqiang Liu, Zhizhou Zhong, Yuan Zhang, Pengfei Wan, and Di Zhang. Livepor- trait: Efficient portrait animation with stitching and retarget- ing control. arXiv preprint arXiv:2407.03168, 2024. 2, 3, 6, 7

    2024 arXiv

  11. [19]

    Mimaface: Face animation via motion-identity modulated appearance feature learning

    Yue Han, Junwei Zhu, Yuxiang Feng, Xiaozhong Ji, Keke He, Xiangtai Li, Yong Liu, et al. Mimaface: Face animation via motion-identity modulated appearance feature learning. arXiv preprint arXiv:2409.15179, 2024. 3

    2024 arXiv

  12. [20]

    Depth-aware generative adversarial network for talking head video generation

    Fa-Ting Hong, Longhao Zhang, Li Shen, and Dan Xu. Depth-aware generative adversarial network for talking head video generation. In Proceedings of the IEEE/CVF con- ference on computer vision and pattern recognition , pages 3397–3406, 2022. 7

    2022

  13. [21]

    Image quality metrics: Psnr vs

    Alain Hore and Djemel Ziou. Image quality metrics: Psnr vs. ssim. In 2010 20th international conference on pattern recognition, pages 2366–2369. IEEE, 2010. 6

    2010

  14. [22]

    Animate anyone: Consistent and controllable image-to-video synthesis for character animation

    Li Hu, Xin Gao, Peng Zhang, Ke Sun, Bang Zhang, and Liefeng Bo. Animate anyone: Consistent and controllable image-to-video synthesis for character animation. arXiv preprint arXiv:2311.17117, 2023. 3

    2023 arXiv

  15. [23]

    You said that?: Synthesising talking faces from audio

    Amir Jamaludin, Joon Son Chung, and Andrew Zisserman. You said that?: Synthesising talking faces from audio. Inter- national Journal of Computer Vision, 127:1767–1779, 2019. 1

    2019

  16. [24]

    Stylevr: Stylizing character animations with normal- izing flows

    Bin Ji, Ye Pan, Yichao Yan, Ruizhao Chen, and Xiaokang Yang. Stylevr: Stylizing character animations with normal- izing flows. IEEE Transactions on Visualization and Com- puter Graphics, 2023. 3

    2023

  17. [25]

    Sport: From zero-shot prompts to real-time motion gener- ation

    Bin Ji, Ye Pan, Zhimeng Liu, Shuai Tan, and Xiaokang Yang. Sport: From zero-shot prompts to real-time motion gener- ation. IEEE Transactions on Visualization and Computer Graphics, 2025. 3

    2025

  18. [26]

    Eamm: One-shot emotional talking face via audio-based emotion-aware motion model

    Xinya Ji, Hang Zhou, Kaisiyuan Wang, Qianyi Wu, Wayne Wu, Feng Xu, and Xun Cao. Eamm: One-shot emotional talking face via audio-based emotion-aware motion model. In ACM SIGGRAPH 2022 Conference Proceedings , pages 1–10, 2022. 6

    2022

  19. [27]

    Loopy: Taming audio-driven portrait avatar with long-term motion dependency

    Jianwen Jiang, Chao Liang, Jiaqi Yang, Gaojie Lin, Tianyun Zhong, and Yanbo Zheng. Loopy: Taming audio-driven portrait avatar with long-term motion dependency. arXiv preprint arXiv:2409.02634, 2024. 3

    2024 arXiv

  20. [28]

    Percep- tual losses for real-time style transfer and super-resolution

    Justin Johnson, Alexandre Alahi, and Li Fei-Fei. Percep- tual losses for real-time style transfer and super-resolution. In Computer Vision–ECCV 2016: 14th European Confer- ence, Amsterdam, The Netherlands, October 11-14, 2016, Proceedings, Part II 14, pages 694–711. Springer, 2016. 4

    2016

  21. [29]

    Blip-2: Bootstrapping language-image pre-training with frozen image encoders and large language models

    Junnan Li, Dongxu Li, Silvio Savarese, and Steven Hoi. Blip-2: Bootstrapping language-image pre-training with frozen image encoders and large language models. In In- ternational conference on machine learning , pages 19730– 19742. PMLR, 2023. 4

    2023

  22. [30]

    Intergen: Diffusion-based multi-human motion genera- tion under complex interactions

    Han Liang, Wenqian Zhang, Wenxuan Li, Jingyi Yu, and Lan Xu. Intergen: Diffusion-based multi-human motion genera- tion under complex interactions. International Journal of Computer Vision, 132(9):3463–3483, 2024. 2

    2024

  23. [31]

    Cyberhost: Taming audio- driven avatar diffusion model with region codebook atten- tion

    Gaojie Lin, Jianwen Jiang, Chao Liang, Tianyun Zhong, Ji- aqi Yang, and Yanbo Zheng. Cyberhost: Taming audio- driven avatar diffusion model with region codebook atten- tion. arXiv preprint arXiv:2409.01876, 2024. 3

    2024 arXiv

  24. [32]

    Anitalker: Animate vivid and di- verse talking faces through identity-decoupled facial motion encoding

    Tao Liu, Feilong Chen, Shuai Fan, Chenpeng Du, Qi Chen, Xie Chen, and Kai Yu. Anitalker: Animate vivid and di- verse talking faces through identity-decoupled facial motion encoding. arXiv preprint arXiv:2405.03121, 2024. 2, 3, 7

    2024 arXiv

  25. [33]

    Vqtalker: Towards multilingual talking avatars through facial motion tokenization

    Tao Liu, Ziyang Ma, Qi Chen, Feilong Chen, Shuai Fan, Xie Chen, and Kai Yu. Vqtalker: Towards multilingual talking avatars through facial motion tokenization. arXiv preprint arXiv:2412.09892, 2024. 2

    2024 arXiv

  26. [34]

    Cvthead: One-shot controllable head avatar with vertex-feature transformer

    Haoyu Ma, Tong Zhang, Shanlin Sun, Xiangyi Yan, Kun Han, and Xiaohui Xie. Cvthead: One-shot controllable head avatar with vertex-feature transformer. In Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision, pages 6131–6141, 2024. 2

    2024

  27. [35]

    Follow-your-emoji: Fine-controllable and expressive freestyle portrait animation

    Yue Ma, Hongyu Liu, Hongfa Wang, Heng Pan, Yingqing He, Junkun Yuan, Ailing Zeng, Chengfei Cai, Heung-Yeung Shum, Wei Liu, et al. Follow-your-emoji: Fine-controllable and expressive freestyle portrait animation. arXiv preprint arXiv:2406.01900, 2024. 2, 3, 6, 7

    2024 arXiv

  28. [36]

    Echomimicv2: Towards striking, simplified, and semi-body human animation

    Rang Meng, Xingyu Zhang, Yuming Li, and Chenguang Ma. Echomimicv2: Towards striking, simplified, and semi-body human animation. arXiv preprint arXiv:2411.10061, 2024. 3, 8

    2024

  29. [37]

    completely blind

    Anish Mittal, Rajiv Soundararajan, and Alan C Bovik. Mak- ing a “completely blind” image quality analyzer. IEEE Sig- nal processing letters, 20(3):209–212, 2012. 6

    2012

  30. [38]

    Emotional voice puppetry

    Ye Pan, Ruisi Zhang, Shengran Cheng, Shuai Tan, Yu Ding, Kenny Mitchell, and Xubo Yang. Emotional voice puppetry. IEEE Transactions on Visualization and Computer Graphics, 29(5):2527–2535, 2023. 3

    2023

  31. [39]

    Expressive talking avatars

    Ye Pan, Shuai Tan, Shengran Cheng, Qunfen Lin, Zijiao Zeng, and Kenny Mitchell. Expressive talking avatars. IEEE Transactions on Visualization and Computer Graph- ics, 2024

    2024

  32. [40]

    Vasa-rig: Audio-driven 3d facial animation with ‘live’mood dynamics in virtual reality

    Ye Pan, Chang Liu, Sicheng Xu, Shuai Tan, and Jiaolong Yang. Vasa-rig: Audio-driven 3d facial animation with ‘live’mood dynamics in virtual reality. IEEE Transactions on Visualization and Computer Graphics, 2025. 3

    2025

  33. [41]

    Dpe: Dis- entanglement of pose and expression for general video por- trait editing

    Youxin Pang, Yong Zhang, Weize Quan, Yanbo Fan, Xi- aodong Cun, Ying Shan, and Dong-ming Yan. Dpe: Dis- entanglement of pose and expression for general video por- trait editing. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages 427–436,

  34. [42]

    Controlnext: Powerful and effi- cient control for image and video generation

    Bohao Peng, Jian Wang, Yuechen Zhang, Wenbo Li, Ming- Chang Yang, and Jiaya Jia. Controlnext: Powerful and effi- cient control for image and video generation. arXiv preprint arXiv:2408.06070, 2024. 3

    2024 arXiv

  35. [43]

    High-resolution image syn- thesis with latent diffusion models

    Robin Rombach, Andreas Blattmann, Dominik Lorenz, Patrick Esser, and Bj¨orn Ommer. High-resolution image syn- thesis with latent diffusion models. In CVPR, pages 10684– 10695, 2022. 3

    2022

  36. [44]

    pytorch-fid: FID Score for PyTorch

    Maximilian Seitzer. pytorch-fid: FID Score for PyTorch. https://github.com/mseitzer/pytorch-fid ,

  37. [45]

    Learning dynamic facial radiance fields for few-shot talking head synthesis

    Shuai Shen, Wanhua Li, Zheng Zhu, Yueqi Duan, Jie Zhou, and Jiwen Lu. Learning dynamic facial radiance fields for few-shot talking head synthesis. In European Conference on Computer Vision, pages 666–682. Springer, 2022. 3

    2022

  38. [46]

    Difftalk: Crafting diffusion models for generalized audio-driven portraits animation

    Shuai Shen, Wenliang Zhao, Zibin Meng, Wanhua Li, Zheng Zhu, Jie Zhou, and Jiwen Lu. Difftalk: Crafting diffusion models for generalized audio-driven portraits animation. In Proceedings of the IEEE/CVF Conference on Computer Vi- sion and Pattern Recognition, pages 1982–1991, 2023. 3

    1982

  39. [47]

    Motionstone: Decoupled motion in- tensity modulation with diffusion transformer for image-to- video generation

    Shuwei Shi, Biao Gong, Xi Chen, Dandan Zheng, Shuai Tan, Zizheng Yang, Yuyuan Li, Jingwen He, Kecheng Zheng, Jingdong Chen, et al. Motionstone: Decoupled motion in- tensity modulation with diffusion transformer for image-to- video generation. CVPR 2025, 2025. 3

    2025

  40. [48]

    Very deep convo- lutional networks for large-scale image recognition

    Karen Simonyan and Andrew Zisserman. Very deep convo- lutional networks for large-scale image recognition. arXiv preprint arXiv:1409.1556, 2014. 4

    2014 arXiv

  41. [49]

    Talking face generation by conditional recurrent adversarial network

    Yang Song, Jingwen Zhu, Dawei Li, Andy Wang, and Hairong Qi. Talking face generation by conditional recurrent adversarial network. In Proceedings of the Twenty-Eighth In- ternational Joint Conference on Artificial Intelligence, 2019. 3

    2019

  42. [50]

    Emmn: Emotional motion memory network for audio-driven emotional talking face generation

    Shuai Tan, Bin Ji, and Ye Pan. Emmn: Emotional motion memory network for audio-driven emotional talking face generation. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 22146–22156, 2023. 3

    2023

  43. [51]

    Say anything with any style

    Shuai Tan, Bin Ji, Yu Ding, and Ye Pan. Say anything with any style. In Proceedings of the AAAI Conference on Artifi- cial Intelligence, pages 5088–5096, 2024

    2024

  44. [52]

    Flowvqtalker: High-quality emotional talking face generation through normalizing flow and quantization

    Shuai Tan, Bin Ji, and Ye Pan. Flowvqtalker: High-quality emotional talking face generation through normalizing flow and quantization. In Proceedings of the IEEE/CVF Con- ference on Computer Vision and Pattern Recognition, pages 26317–26327, 2024. 7

    2024

  45. [53]

    Style2talker: High-resolution talking head generation with emotion style and art style

    Shuai Tan, Bin Ji, and Ye Pan. Style2talker: High-resolution talking head generation with emotion style and art style. In Proceedings of the AAAI Conference on Artificial Intelli- gence, pages 5079–5087, 2024. 3

    2024

  46. [54]

    Mimir: Improving video diffusion models for precise text understanding

    Shuai Tan, Biao Gong, Yutong Feng, Kecheng Zheng, Dan- dan Zheng, Shuwei Shi, Yujun Shen, Jingdong Chen, and Ming Yang. Mimir: Improving video diffusion models for precise text understanding. In Proceedings of the Computer Vision and Pattern Recognition Conference , pages 2397...

    2025

  47. [55]

    Animate-x: Universal character image ani- mation with enhanced motion representation

    Shuai Tan, Biao Gong, Xiang Wang, Shiwei Zhang, Dandan Zheng, Ruobing Zheng, Kecheng Zheng, Jingdong Chen, and Ming Yang. Animate-x: Universal character image ani- mation with enhanced motion representation. In ICLR 2025,

    2025

  48. [56]

    Synmotion: Semantic-visual adaptation for motion customized video generation

    Shuai Tan, Biao Gong, Yujie Wei, Shiwei Zhang, Zhuoxin Liu, Dandan Zheng, Jingdong Chen, Yan Wang, Hao Ouyang, Kecheng Zheng, and Yujun Shen. Synmotion: Semantic-visual adaptation for motion customized video generation. arXiv preprint arXiv:2506.23690, 2025. 2

    2025 arXiv

  49. [57]

    Edtalk: Effi- cient disentanglement for emotional talking head synthesis

    Shuai Tan, Bin Ji, Mengxiao Bi, and Ye Pan. Edtalk: Effi- cient disentanglement for emotional talking head synthesis. In European Conference on Computer Vision , pages 398–

  50. [58]

    Svp: Style- enhanced vivid portrait talking head diffusion model

    Weipeng Tan, Chuming Lin, Chengming Xu, Xiaozhong Ji, Junwei Zhu, Chengjie Wang, and Yanwei Fu. Svp: Style- enhanced vivid portrait talking head diffusion model. arXiv preprint arXiv:2409.03270, 2024. 3

    2024 arXiv

  51. [59]

    Neural voice puppetry: Audio-driven facial reenactment

    Justus Thies, Mohamed Elgharib, Ayush Tewari, Christian Theobalt, and Matthias Nießner. Neural voice puppetry: Audio-driven facial reenactment. InComputer Vision–ECCV 2020: 16th European Conference, Glasgow, UK, August 23–28, 2020, Proceedings, Part XVI 16 , pages 716–731. Spr...

    2020

  52. [60]

    Emo: Emote portrait alive-generating expressive portrait videos with audio2video diffusion model under weak conditions

    Linrui Tian, Qi Wang, Bang Zhang, and Liefeng Bo. Emo: Emote portrait alive-generating expressive portrait videos with audio2video diffusion model under weak conditions. arXiv preprint arXiv:2402.17485, 2024. 3

    2024 arXiv

  53. [61]

    Musepose: a pose-driven image-to-video framework for virtual human generation

    Zhengyan Tong, Chao Li, Zhaokang Chen, Bin Wu, and Wenjiang Zhou. Musepose: a pose-driven image-to-video framework for virtual human generation. arxiv, 2024. 3

    2024

  54. [62]

    Nonlinear 3d face morphable model

    Luan Tran and Xiaoming Liu. Nonlinear 3d face morphable model. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 7346–7355, 2018. 2

    2018

  55. [63]

    V-express: Conditional dropout for progres- sive training of portrait video generation

    Cong Wang, Kuan Tian, Jun Zhang, Yonghang Guan, Feng Luo, Fei Shen, Zhiwei Jiang, Qing Gu, Xiao Han, and Wei Yang. V-express: Conditional dropout for progres- sive training of portrait video generation. arXiv preprint arXiv:2406.02511, 2024. 3

    2024 arXiv

  56. [64]

    Progressive disentangled representation learning for fine-grained controllable talking head synthesis

    Duomin Wang, Yu Deng, Zixin Yin, Heung-Yeung Shum, and Baoyuan Wang. Progressive disentangled representation learning for fine-grained controllable talking head synthesis. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 17979–17989, ...

    2023

  57. [65]

    Seeing what you said: Talking face gen- eration guided by a lip reading expert

    Jiadong Wang, Xinyuan Qian, Malu Zhang, Robby T Tan, and Haizhou Li. Seeing what you said: Talking face gen- eration guided by a lip reading expert. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 14653–14662, 2023. 7

    2023

  58. [66]

    Lipformer: High- fidelity and generalizable talking face generation with a pre- learned facial codebook

    Jiayu Wang, Kang Zhao, Shiwei Zhang, Yingya Zhang, Yu- jun Shen, Deli Zhao, and Jingren Zhou. Lipformer: High- fidelity and generalizable talking face generation with a pre- learned facial codebook. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recog...

    2023

  59. [67]

    Mead: A large-scale audio-visual dataset for emo- tional talking-face generation

    Kaisiyuan Wang, Qianyi Wu, Linsen Song, Zhuoqian Yang, Wayne Wu, Chen Qian, Ran He, Yu Qiao, and Chen Change Loy. Mead: A large-scale audio-visual dataset for emo- tional talking-face generation. In Computer Vision–ECCV 2020: 16th European Conference, Glasgow, UK, August 23– 2...

    2020

  60. [68]

    Audio2head: Audio-driven one-shot talking-head generation with natural head motion

    S Wang, L Li, Y Ding, C Fan, and X Yu. Audio2head: Audio-driven one-shot talking-head generation with natural head motion. In International Joint Conference on Artificial Intelligence. IJCAI, 2021. 1, 3

    2021

  61. [69]

    Disco: Disentangled control for referring human dance generation in real world

    Tan Wang, Linjie Li, Kevin Lin, Chung-Ching Lin, Zhengyuan Yang, Hanwang Zhang, Zicheng Liu, and Li- juan Wang. Disco: Disentangled control for referring human dance generation in real world. In ICLR, 2024. 3

    2024

  62. [70]

    Unianimate: Taming unified video diffusion mod- els for consistent human image animation

    Xiang Wang, Shiwei Zhang, Changxin Gao, Jiayu Wang, Xiaoqiang Zhou, Yingya Zhang, Luxin Yan, and Nong Sang. Unianimate: Taming unified video diffusion mod- els for consistent human image animation. arXiv preprint arXiv:2406.01188, 2024. 3

    2024 arXiv

  63. [71]

    Latent image animator: Learning to animate im- ages via latent space navigation

    Yaohui Wang, Di Yang, Francois Bremond, and Antitza Dantcheva. Latent image animator: Learning to animate im- ages via latent space navigation. InInternational Conference on Learning Representations, 2021. 2, 8

    2021

  64. [72]

    Image quality assessment: from error visibility to structural similarity

    Zhou Wang, Alan C Bovik, Hamid R Sheikh, and Eero P Si- moncelli. Image quality assessment: from error visibility to structural similarity. IEEE Transactions on Image Process- ing, 13(4):600–612, 2004. 6

    2004

  65. [73]

    Restoreformer: High-quality blind face restoration from undegraded key-value pairs

    Zhouxia Wang, Jiawei Zhang, Runjian Chen, Wenping Wang, and Ping Luo. Restoreformer: High-quality blind face restoration from undegraded key-value pairs. InProceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 17512–17521, 2022. 6

    2022

  66. [74]

    Aniportrait: Audio-driven synthesis of photorealistic portrait animation

    Huawei Wei, Zejun Yang, and Zhisheng Wang. Aniportrait: Audio-driven synthesis of photorealistic portrait animation. arXiv preprint arXiv:2403.17694, 2024. 3

    2024 arXiv

  67. [75]

    F 3a-gan: Facial flow for face animation with generative adversarial networks

    Xintian Wu, Qihang Zhang, Yiming Wu, Huanyu Wang, Songyuan Li, Lingyun Sun, and Xi Li. F 3a-gan: Facial flow for face animation with generative adversarial networks. IEEE Transactions on Image Processing , 30:8658–8670,

  68. [76]

    Vfhq: A high-quality dataset and bench- mark for video face super-resolution

    Liangbin Xie, Xintao Wang, Honglun Zhang, Chao Dong, and Ying Shan. Vfhq: A high-quality dataset and bench- mark for video face super-resolution. In The IEEE Con- ference on Computer Vision and Pattern Recognition Work- shops (CVPRW), 2022. 6

    2022

  69. [77]

    X-portrait: Expressive portrait anima- tion with hierarchical motion attention

    You Xie, Hongyi Xu, Guoxian Song, Chao Wang, Yichun Shi, and Linjie Luo. X-portrait: Expressive portrait anima- tion with hierarchical motion attention. In ACM SIGGRAPH 2024 Conference Papers, pages 1–11, 2024. 2, 3, 6, 7

    2024

  70. [78]

    Hallo: Hierarchical audio-driven vi- sual synthesis for portrait image animation

    Mingwang Xu, Hui Li, Qingkun Su, Hanlin Shang, Li- wei Zhang, Ce Liu, Jingdong Wang, Luc Van Gool, Yao Yao, and Siyu Zhu. Hallo: Hierarchical audio-driven vi- sual synthesis for portrait image animation. arXiv preprint arXiv:2406.08801, 2024. 2, 3, 7, 8

    2024 arXiv

  71. [79]

    Vasa-1: Lifelike audio-driven talking faces generated in real time

    Sicheng Xu, Guojun Chen, Yu-Xiao Guo, Jiaolong Yang, Chong Li, Zhenyu Zang, Yizhong Zhang, Xin Tong, and Baining Guo. Vasa-1: Lifelike audio-driven talking faces generated in real time. arXiv preprint arXiv:2404.10667 ,

    arXiv

  72. [80]

    Magicanimate: Temporally consistent human im- age animation using diffusion model

    Zhongcong Xu, Jianfeng Zhang, Jun Hao Liew, Hanshu Yan, Jia-Wei Liu, Chenxu Zhang, Jiashi Feng, and Mike Zheng Shou. Magicanimate: Temporally consistent human im- age animation using diffusion model. arXiv preprint arXiv:2311.16498, 2023. 3

    2023 arXiv

  73. [81]

    Smgdiff: Soccer motion generation using diffusion probabilistic mod- els

    Hongdi Yang, Chengyang Li, Zhenxuan Wu, Gaozheng Li, Jingya Wang, Jingyi Yu, Zhuo Su, and Lan Xu. Smgdiff: Soccer motion generation using diffusion probabilistic mod- els. arXiv preprint arXiv:2411.16216, 2024. 2

    2024 arXiv

  74. [82]

    Face2face ρ: Real- time high-resolution one-shot face reenactment

    Kewei Yang, Kang Chen, Daoliang Guo, Song-Hai Zhang, Yuan-Chen Guo, and Weidong Zhang. Face2face ρ: Real- time high-resolution one-shot face reenactment. InEuropean conference on computer vision, pages 55–71. Springer, 2022. 3

    2022

  75. [83]

    Megactor: Harness the power of raw video for vivid portrait animation

    Shurong Yang, Huadong Li, Juhao Wu, Minhao Jing, Linze Li, Renhe Ji, Jiajun Liang, and Haoqiang Fan. Megactor: Harness the power of raw video for vivid portrait animation. arXiv preprint arXiv:2405.20851, 2024. 3

    2024 arXiv

  76. [84]

    Superb: Speech processing universal performance benchmark.arXiv preprint arXiv:2105.01051, 2021

    Shu-wen Yang, Po-Han Chi, Yung-Sung Chuang, Cheng- I Jeff Lai, Kushal Lakhotia, Yist Y Lin, Andy T Liu, Jiatong Shi, Xuankai Chang, Guan-Ting Lin, et al. Superb: Speech processing universal performance benchmark.arXiv preprint arXiv:2105.01051, 2021. 5

    2021 arXiv

  77. [85]

    Mesh guided one-shot face reenactment using graph convo- lutional networks

    Guangming Yao, Yi Yuan, Tianjia Shao, and Kun Zhou. Mesh guided one-shot face reenactment using graph convo- lutional networks. In Proceedings of the 28th ACM inter- national conference on multimedia, pages 1773–1781, 2020. 2

    2020

  78. [86]

    Few-shot adversarial learning of realistic neural talking head models

    Egor Zakharov, Aliaksandra Shysheya, Egor Burkov, and Victor Lempitsky. Few-shot adversarial learning of realistic neural talking head models. InProceedings of the IEEE/CVF international conference on computer vision , pages 9459– 9468, 2019. 3

    2019

  79. [87]

    Metaportrait: Identity-preserving talking head gener- ation with fast personalized adaptation

    Bowen Zhang, Chenyang Qi, Pan Zhang, Bo Zhang, Hsiang- Tao Wu, Dong Chen, Qifeng Chen, Yong Wang, and Fang Wen. Metaportrait: Identity-preserving talking head gener- ation with fast personalized adaptation. In Proceedings of the IEEE/CVF Conference on Computer Vision and Patte...

    2023

  80. [88]

    The unreasonable effectiveness of deep features as a perceptual metric

    Richard Zhang, Phillip Isola, Alexei A Efros, Eli Shecht- man, and Oliver Wang. The unreasonable effectiveness of deep features as a perceptual metric. In Proceedings of the IEEE conference on computer vision and pattern recogni- tion, pages 586–595, 2018. 4

    2018

  81. [89]

    Sadtalker: Learning realistic 3d motion coefficients for stylized audio- driven single image talking face animation

    Wenxuan Zhang, Xiaodong Cun, Xuan Wang, Yong Zhang, Xi Shen, Yu Guo, Ying Shan, and Fei Wang. Sadtalker: Learning realistic 3d motion coefficients for stylized audio- driven single image talking face animation. In Proceedings of the IEEE/CVF Conference on Computer Vision and P...

    2023

  82. [90]

    Mim- icmotion: High-quality human motion video generation with confidence-aware pose guidance

    Yuang Zhang, Jiaxi Gu, Li-Wen Wang, Han Wang, Junqi Cheng, Yuefeng Zhu, and Fangyuan Zou. Mim- icmotion: High-quality human motion video generation with confidence-aware pose guidance. arXiv preprint arXiv:2406.19680, 2024. 3

    2024 arXiv

  83. [91]

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

    Zhimeng Zhang, Lincheng Li, Yu Ding, and Changjie Fan. Flow-guided one-shot talking face generation with a high-resolution audio-visual dataset. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 3661–3670, 2021. 6

    2021

  84. [92]

    Thin-plate spline motion model for image animation

    Jian Zhao and Hui Zhang. Thin-plate spline motion model for image animation. In Proceedings of the IEEE/CVF Con- ference on Computer Vision and Pattern Recognition, pages 3657–3666, 2022. 2

    2022

  85. [93]

    Media2face: Co-speech facial animation gen- eration with multi-modality guidance

    Qingcheng Zhao, Pengyu Long, Qixuan Zhang, Dafei Qin, Han Liang, Longwen Zhang, Yingliang Zhang, Jingyi Yu, and Lan Xu. Media2face: Co-speech facial animation gen- eration with multi-modality guidance. In ACM SIGGRAPH 2024 conference papers, pages 1–13, 2024

    2024

  86. [94]

    Sparse to dense motion transfer for face image animation

    Ruiqi Zhao, Tianyi Wu, and Guodong Guo. Sparse to dense motion transfer for face image animation. In Proceedings of the IEEE/CVF International Conference on Computer Vi- sion, pages 1991–2000, 2021. 2

    1991

  87. [95]

    Identity- preserving talking face generation with landmark and ap- pearance priors

    Weizhi Zhong, Chaowei Fang, Yinqi Cai, Pengxu Wei, Gangming Zhao, Liang Lin, and Guanbin Li. Identity- preserving talking face generation with landmark and ap- pearance priors. In Proceedings of the IEEE/CVF Confer- ence on Computer Vision and Pattern Recognition , pages 9729–...

    2023

  88. [96]

    Talking face generation by adversarially disentangled audio-visual representation

    Hang Zhou, Yu Liu, Ziwei Liu, Ping Luo, and Xiaogang Wang. Talking face generation by adversarially disentangled audio-visual representation. In Proceedings of the AAAI con- ference on artificial intelligence, pages 9299–9306, 2019. 1, 3

    2019

  89. [97]

    Makelttalk: speaker-aware talking-head animation

    Yang Zhou, Xintong Han, Eli Shechtman, Jose Echevar- ria, Evangelos Kalogerakis, and Dingzeyu Li. Makelttalk: speaker-aware talking-head animation. ACM Transactions On Graphics (TOG), 39(6):1–15, 2020. 3

    2020

  90. [98]

    Champ: Controllable and consistent human image animation with 3d parametric guidance

    Shenhao Zhu, Junming Leo Chen, Zuozhuo Dai, Yinghui Xu, Xun Cao, Yao Yao, Hao Zhu, and Siyu Zhu. Champ: Controllable and consistent human image animation with 3d parametric guidance. In European Conference on Computer Vision (ECCV), 2024. 3

    2024

  91. [99]

    Face alignment in full pose range: A 3d total solution.IEEE trans- actions on pattern analysis and machine intelligence, 41(1): 78–92, 2017

    Xiangyu Zhu, Xiaoming Liu, Zhen Lei, and Stan Z Li. Face alignment in full pose range: A 3d total solution.IEEE trans- actions on pattern analysis and machine intelligence, 41(1): 78–92, 2017. 2

    2017

  92. [416]

    Springer, 2025. 2, 3, 7

    2025

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.