Pith. sign in

REVIEW 4 major objections 5 minor 184 references

Multi-View Face and Gesture Animation with Dynamic Gaussians

T0 review · 4 major / 5 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 introduces MVFGA, a multi-view pipeline that reconstructs a photorealistic upper-body avatar from synchronized cameras and re-animates it with fine-grained facial expressions and hand gestures, reporting the best image quality…

desk verdict Solid systems paper with a useful dataset, but the claim that it outperforms baselines is not backed by its own evaluation because the baselines get one source frame while MVFGA trains on 17 synchronized cameras. read the letter →

arxiv 2608.04722 v1 pith:KMIZZUPE submitted 2026-08-05 cs.CV cs.GR

Alireza Javanmardi , Vippin Kumar Jeetmal , Christen Millerdurai , Alain Pagani , Didier Stricker This is my paper
classification cs.CVcs.GR
keywords multi-viewreconstruction3DGaussiansplattingupper-bodyavatarfacialanimationhandgesturesynthesisparametricbodymodelnovel-viewrenderingmotioncapturedataset
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

This paper introduces MVFGA, a pipeline that builds a photorealistic, animatable 3D upper-body avatar from synchronized multi-view video and re-drives it with novel facial expressions, hand gestures, and body poses. The central claim is that explicitly modeling the face and hands with dedicated parametric models and fusing them into an upper-body SMPL-X mesh, then attaching 3D Gaussians to the mesh, gives sharper facial detail and more accurate hand articulation than existing generative or full-body avatar methods. On a new 17-camera, 15-subject dataset, the paper reports the best scores across all compared baselines on image quality, pose accuracy, and identity preservation (e.g., LPIPS 0.053, PSNR 25.93, AKD_H 0.71, CSIM 0.85). If the claim holds, upper-body avatars can convey the subtle expression and gesture cues that matter for natural remote communication while rendering in real time (about 49 FPS).

What carries the argument

The load-bearing mechanism is the surface-attached 3D Gaussian splat: each triangle of the fused upper-body mesh carries at least one Gaussian stored in the triangle's local coordinate frame and mapped to global space by that triangle's rotation, translation, and scale. This binding, adapted from existing per-triangle Gaussian initialization, lets the appearance field deform with the mesh, while a multi-view candidate-selection step supplies the parametric geometry by generating several monocular FLAME and MANO fits and keeping the one with the lowest aggregated 2D reprojection error. The optimization objective combines an RGB reconstruction loss with a perceptual LPIPS term plus position and scaling regularizers that keep splats near their parent triangles, and an adaptive density control adds or removes splats while always retaining at least one per triangle.

What would settle it

A concrete test is to run MVFGA on a captured sequence with head yaw beyond ±50 degrees and a gesture with heavily occluded fingers, then compare hand keypoint error (AKD_H) and identity similarity (CSIM) against the ablation that removes FLAME and MANO; if that ablation matches or beats the full method, or if MVFGA's hand AKD rises above the GUAVA baseline on such frames, the claim that explicit face and hand modeling is required for high-fidelity upper-body animation would be contradicted.

Watch

Extended reading notes

Core claim

The paper's central claim is that fine-grained facial expressions and articulated hand motions can be preserved in a multi-view-consistent 3D avatar if the parametric upper-body model is extended to accept separately fitted FLAME and MANO parameters, and if the appearance is rendered by 3D Gaussians that are bound to the mesh triangles and optimized jointly with a light fine-tuning of the body parameters. The evaluation on the introduced MVFGA-MoCap dataset shows MVFGA achieving the best L1 error (4.06), SSIM (0.938), PSNR (25.93), LPIPS (0.053), hand and torso keypoint distance, and identity cosine similarity (0.85) among the compared generative and Gaussian-based baselines, and the ablations indicate that removing FLAME/MANO, upper-body fine-tuning, or LPIPS supervision each degrades quality. Qualitatively, the method preserves finger structure under strong self-occlusion and maintains target identity in cross-identity motion transfer, while novel-view renderings stay consistent at yaw offsets up to 20 degrees.

Load-bearing premise

The pipeline depends on the monocular face and hand estimators being accurate enough after multi-view candidate selection; if their fitted mesh parameters are systematically wrong for a pose or expression, the Gaussian optimization can only refine appearance, not fix the geometry, so the avatar's face and hands will drift from the intended motion.

Editorial extensions

If this is right

  • Upper-body telepresence and VR avatars could show natural communicative cues such as subtle smiles, raised brows, and precise finger gestures instead of a static face with generic hands, improving comfort in remote interaction.
  • The explicit face and hand parametric fusion can be reused as a component in other mesh-based avatar pipelines, because the resulting upper-body model retains SMPL-X compatibility.
  • The MVFGA-MoCap dataset provides a multi-view benchmark with controlled expressions and gestures, enabling direct comparisons for future upper-body reenactment methods.
  • Real-time rendering at about 49 FPS on a single RTX 3090 indicates the representation is suitable for interactive avatar applications, not just offline reconstruction.

Reading between the lines

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

  • Because the parametric geometry comes from monocular estimators with a multi-view selection heuristic, a natural extension is to train the face and hand regressors directly on the multi-view captures or add temporal smoothing, which would likely harden the pipeline against extreme poses without changing the Gaussian rendering core.
  • The per-triangle Gaussian binding could transfer to whole-body avatars, but the paper's stated limitation on secondary motion (clothing deformation) means a separate dynamic layer would still be needed for realistic garments.
  • The dataset is recorded under studio lighting with a 150-degree camera arc, so the reported photorealistic quality is best read as holding under those controlled capture conditions; an in-the-wild monocular application would need additional regularization or a different estimator.
  • The multi-view reprojection-based candidate selection could also be applied to body pose estimation, potentially improving torso accuracy in the same framework without additional training data.
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 / 5 minor

Summary. MVFGA is a multi-view pipeline for upper-body avatar reconstruction and reenactment. The method first fits a custom upper-body SMPL-X model to synchronized 17-camera video, injecting FLAME face parameters from MICA and MANO hand parameters from HaMeR, and selects consistent estimates via multi-view reprojection. A triangle-bound 3D Gaussian representation is then optimized with L1/D-SSIM/LPIPS losses plus regularization and adaptive density control. The paper also introduces MVFGA-MoCap, a 15-subject multi-view dataset with controlled facial expressions and hand gestures. Quantitative evaluation on four held-out identities is reported for self-reenactment (image quality, keypoint distance, identity similarity), novel-view synthesis is shown qualitatively, and ablations on one subject examine the contributions of face/hand modeling, parametric fine-tuning, and LPIPS supervision. The central claim is that MVFGA outperforms prior generative and graphics-based baselines for upper-body avatar animation.

Significance. If the method and dataset were validated independently, the work would be a useful contribution: it combines per-identity multi-view appearance optimization with explicit face/hand parametric control, reports a high rendering speed (about 49 FPS on an RTX 3090), and introduces a new multi-view upper-body dataset. The strengths are the clear pipeline, a sensible decomposition of face, hands, and body, ablations showing that each proposed component matters, and an honest limitations section. However, the evidence presented here does not yet support the headline comparative claim: the comparisons in Tables 1 and 2 are against baselines operating under a different information regime, the pose metric is partly circular, and the statistical basis is thin. The dataset and the temporal-consistency metric come from the authors' own prior work, so external validation is still missing. The paper does not mention code release, which also limits reproducibility.

major comments (4)
  1. [Section 4.3, Section 4.5, Tables 1 and 2] Tables 1 and 2 do not support the abstract's 'outperforming baselines for upper-body avatar animation' claim as stated. MVFGA is trained end-to-end on the target identity from 17 calibrated cameras for 600K iterations (Section 4.2), while each baseline is applied in the standard one-shot setting, receiving only the first frame of the sequence as source and the remaining frames as driving video (Section 4.5). This is an apples-to-oranges comparison: the large margins (e.g., L1 4.06 vs 9.59, PSNR 25.93 vs 19.62, AKD_H 0.71 vs 1.98) are plausibly explained by the additional target-specific multi-view training data rather than by the method's representational choices. Please either explicitly scope the claim to 'given multi-view training of the target identity' or add a same-information comparison, e.g., fine-tuning the baselines on the same multi-view data or evaluating MVFGA in a one-shot/few-shot mode.
  2. [Section 4.4, Section 4.2, Appendix A] Table 2 uses Average Keypoint Distance computed on MediaPipe landmarks (Section 4.4) to measure face, hand, and torso pose accuracy. The same MediaPipe holistic keypoints are an explicit observation source for the SMPL-X fitting that produces the animated mesh (Section 4.2; Appendix A, 'Pose Estimation' and 'SMPL-X Parameters'). The low AKD values in Table 2 are therefore at least partly a measure of how well the system fits the evaluation detector, not an independent measure of pose accuracy. Please report pose errors against an independent reference, such as manual annotations, triangulated keypoints from held-out views, or a different detector that is not used during fitting, and state which keypoints were used in the mesh fitting.
  3. [Section 4.5, Section 4.6, Tables 1–3] All headline numbers in Tables 1 and 2 are single scalars aggregated over only four held-out identities (Section 4.5), and the ablation study is performed on a single held-out subject (Section 4.6, Table 3). No variance, confidence intervals, or significance tests are reported. Differences such as AKD_T 1.37 (Ours) vs 1.39 (GUAVA) or AKD_F 0.18 vs 0.15 are likely within noise, so the statement that the method achieves 'the best overall performance across all metrics' is not statistically supported. Please report per-identity or per-clip distributions and compute paired significance tests across identities.
  4. [Section 4.5.2, Figure 7, Table 3] Novel-view synthesis, a central claimed capability in the abstract and introduction ('multi-view-consistent', 'novel viewpoints'), is evaluated only qualitatively against GUAVA in Figure 7, and the novel-view columns of Table 3 are ablations of the proposed method without any baseline. This is insufficient evidence for the novel-view contribution. Please add quantitative novel-view metrics (PSNR/SSIM/LPIPS and keypoint errors) for all methods that support view extrapolation, across the held-out identities and for a range of yaw/pitch angles, and specify which cameras are held out.
minor comments (5)
  1. [Section 4.4, Figure 9] Temporal Jittering Error (TJE) is referenced by citation [JJH*26] but is not defined in the paper; please include its formula and the alignment convention used in Figure 9, especially the meaning of Δ=2.
  2. [Section 3.1] The statement 'we generate multiple candidate estimates and select the one with the lowest aggregated multi-view 2D reprojection error' does not specify how candidates are generated, how many are considered, or how outliers are rejected; please provide these details in the main text or the supplementary material.
  3. [Section 3.4 and Section 4.5.2] The dataset has 15 front-facing cameras spanning 150 degrees and two rear cameras used only for body keypoint detection, and Section 4.8 acknowledges that full 360-degree rendering remains challenging; please consistently say 'frontal novel views' rather than general 'novel viewpoints'.
  4. [Table 1 caption and Table 3 caption] The numbers for 'Ours' differ between Table 1 (L1 4.06, self-reenactment) and Table 3 (L1 2.71, self-reenactment) because they are on different subjects or settings; please state this explicitly in the captions to avoid confusion.
  5. [Section 4.8] The stated limitations (difficulty of obtaining the parametric model, no full 360-degree rendering, no secondary motion) are relevant to the framing: the introduction and conclusion currently present the parametric model and multi-view consistency as strengths without these caveats, so please align the claims with the acknowledged limitations.

Circularity Check

1 steps flagged · score 6.0 of 10

Pose-accuracy claim in Table 2 is partially circular: AKD is computed on the same MediaPipe landmarks used to fit MVFGA's driving mesh; image-quality claims remain independent.

  1. fitted input called prediction [Sec. 4.2 and Sec. 4.4 (Table 2); Appendix A 'SMPL-X Parameters', Eq. (S10)]
    "Initial 2D keypoints from MediaPipe [LTN*19] are combined with detailed FLAME [LBB*17] face parameters (from MICA [ZBT22]) and MANO [RTB17] hand parameters (from HaMeR [PSR*24]) within the EasyMocap [DFJ*21] framework to produce temporally consistent SMPL-X [PCG*19] parameters. ... For pose accuracy, we report Average Keypoint Distance (AKD) [Gas11], computed on MediaPipe landmarks [LTN*19] for the face, hands, and torso."

    The mesh that drives MVFGA's self-reenactment renderings is obtained by the Sec. 3.1 fitting: SMPL-X parameters are optimized against multi-view 2D keypoints, and EasyMocap's SMPL-X objective (Eq. S10) contains the 2D re-projection term lambda_2D L_2D over the MediaPipe keypoints. AKD in Table 2 is computed on the same MediaPipe landmarks. Hence a low AKD for MVFGA is partly guaranteed by construction: the rendered landmarks are projections of a mesh fit to those exact detections, so the metric measures fitting consistency rather than animation generalization. The effect is strongest for hands and torso, which are the headline pose gains, and the face is also affected because MediaPipe face landmarks enter the final SMPL-X fusion.

full rationale

Most of the derivation is self-contained and not circular. The avatar geometry is produced by external monocular estimators (MICA, HaMeR) and a multi-view SMPL-X fit; appearance is optimized by 3D Gaussian splatting against real multi-view images with RGB, SSIM, and LPIPS losses (Eq. 7) on held-out frames, so the L1/PSNR/SSIM/LPIPS results are genuine. Ablations compare variants of the same pipeline. The authors' own dataset and the TJE metric [JJH*26] are self-produced but not load-bearing circularity: TJE is a generic temporal metric applied uniformly to all methods, and the dataset is a new contribution with held-out identities and frames. The one concrete circularity is the AKD evaluation: the same MediaPipe landmarks that drive the mesh fit are reused as ground truth for pose accuracy, so Table 2's AKD gains are partly forced. The comparison asymmetry (MVFGA trained on multi-view target identity versus one-shot baselines) is a serious evaluation confound, but it is not circularity under the definitions used here. The limitations section acknowledges parametric-model reliance and frontal-view restriction, which does not reveal additional circular steps. Overall, there is partial circularity in one pose metric, while the central rendering claim retains independent content.

Assumptions & free parameters 6 free parameters · 5 assumptions · 0 invented entities

The pipeline introduces no new theoretical entities. It relies on established parametric human models (SMPL-X, FLAME, MANO) and pre-trained estimators (MICA, HaMeR) as domain assumptions. The listed hyperparameters are set by hand and affect optimization, but they are not fitted to the evaluation targets.

free parameters (6)
  • loss weight λ (L1 vs D-SSIM) = 0.2
    Set by hand in Eq. 7 and Section 4.2; affects reconstruction quality.
  • loss weight γ (LPIPS) = 0.04
    From Eq. 7, Section 4.2.
  • loss weight λ_position = 0.01
    Regularizes Gaussian mean positions, Eq. 8, Section 4.2.
  • loss weight λ_scaling = 1
    Regularizes Gaussian scaling, Eq. 9, Section 4.2.
  • position regularization threshold ϵ_position = 1
    Allows small deviations from parent triangle, Eq. 8.
  • scaling regularization threshold ϵ_scaling = 0.6
    Disables penalty for small splats, Eq. 9.
assumptions (5)
  • domain assumption The parametric human models SMPL-X, FLAME, and MANO provide faithful representations of body, face, and hand shape and articulation.
    The entire reconstruction and animation pipeline is built on these models; invoked in Section 3.1 and the Supplementary Appendix.
  • domain assumption MICA and HaMeR provide accurate monocular face and hand parameter estimates.
    These pre-trained estimators initialize the face and hand parameters; Section 3.1.
  • domain assumption 3D Gaussians attached to mesh triangles deform correctly under the articulated motion defined by the parametric model.
    Follows the GaussianAvatar formulation (Section 3.3); if the mesh is wrong, the attached Gaussians inherit the error.
  • domain assumption Background matting (BiRefNet) and semantic body-part segmentation (Sapiens) correctly isolate the upper body and produce reliable masks.
    The masks are used both to supervise Gaussian optimization and to evaluate the rendered avatar; Section 3.1 and 4.2.
  • domain assumption The camera calibration (intrinsics and extrinsics) is accurate enough to support multi-view reprojection and triangulation.
    Multi-view consistency, candidate selection, and SMPL-X fitting all depend on calibrated cameras; Appendix A.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Multi-View Face and Gesture Animation with Dynamic Gaussians." pith.science (2026). https://pith.science/paper/KMIZZUPE

@misc{pith2026260804722,
  author       = {Pith},
  title        = {Pith review of: Multi-View Face and Gesture Animation with Dynamic Gaussians},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/KMIZZUPE}},
  note         = {Machine review of arXiv:2608.04722}
}
read the original abstract

Creating photorealistic 3D human avatars with realistic upper-body motion remains challenging. Existing approaches either focus on the head and overlook hand gestures, or reconstruct the full body but fail to preserve fine-grained facial fidelity and hand pose accuracy. As a result, current methods struggle to capture the subtle dynamics of facial expressions and hand gestures that are crucial for natural human communication. While methods based on full-body parametric models enable avatar reconstruction from monocular or multi-view inputs, they often lack accurate facial animation and detailed hand articulation. To address these limitations, we propose MVFGA, a novel multi-view-consistent pipeline for generating realistic upper-body avatars. Our approach models the face and hands separately and fuses them with a parametric upper-body mesh model, enabling the capture of fine-grained facial expressions and hand poses for accurate upper-body avatar reconstruction. We then splat 3D Gaussians onto the obtained mesh, enabling high-quality rendering of dynamic avatars from novel viewpoints. Furthermore, we introduce MVFGA-MoCap, a multi-view upper-body motion capture dataset featuring controlled facial expression sequences, diverse hand gestures, and free-form communication. Experiments show that MVFGA generates visually realistic avatars with high-fidelity facial expressions and hand motions, outperforming baselines for upper-body avatar animation. Project page: https://dfki-av.github.io/MVFGA/

Figures

Figures reproduced from arXiv: 2608.04722 by the authors.

Figure 1
Figure 1. Multi-view face and gesture animation. Given a time-varying driving mesh (left), our method (MVFGA) animates a reconstructed 3D avatar learned from multi-view data and synthesizes geometry-consistent novel-view renderings (right), faithfully reproducing facial expressions and hand gestures. Abstract Creating photorealistic 3D human avatars with realistic upper-body motion remains challenging. Existing approaches eit… view at source ↗
Figure 2
Figure 2. Upper-body mesh generation pipeline: We perform keypoint extraction, body-part segmentation, and background matting, fol￾lowed by refined face and hand fitting. The resulting face, hand, and body parameters are fused to obtain a complete upper-body representa￾tion with accurate shape, pose, and global translation. 2.2. Generative Avatar Animation Generative-based avatar animation methods aim to synthesize ani￾mated … view at source ↗
Figure 3
Figure 3. Overview of our avatar synthesis pipeline. Given multi-view images and corresponding upper-body parameters, following Gaus￾sianAvatar [QKS*24] the template mesh is posed into the deformed space. Each triangle is then assigned a 3D Gaussian representation. These Gaussians are rasterized using a tile-based rasterizer to produce rendered images, which are supervised using an RGB reconstruction loss. An adaptive density… view at source ↗
Figures from the paper (6 more)
Figure 4
Figure 4. Figure 4: Upper-body mesh fitting. From left to right: reference, rendered image, and rendered fitted mesh. 3.2. 3D Gaussian Splatting Preliminary We build on 3D Gaussian Splatting (3DGS) [KKLD23], which re￾constructs a scene from multi-view images and calibrated cameras using a…
Figure 5
Figure 5. Figure 5: Qualitative comparison. Our method captures facial expressions, hand gestures, and upper-body poses across diverse mo￾tions while better preserving appearance compared to MagicAnimate[XZL*24], AnimateAnyone [Hu24], Champ [ZCD*24], and GUAVA [ZLL*25]. MagicAnimate Anima…
Figure 6
Figure 6. Figure 6: Qualitative comparison under complex hand articulation. The highlighted region contains a challenging gesture with closely interacting fingers and strong self-occlusions. Compared with prior methods, our approach better preserves finger structure, hand pose accuracy, a…
Figure 7
Figure 7. Figure 7: Novel-view synthesis. Rendering at unseen yaw angles (+5 ◦ , +20◦ ). Our method preserves geometry and appearance better than GUAVA [ZLL*25] [PITH_FULL_IMAGE:figures/full_fig_p008_7.png]
Figure 9
Figure 9. Figure 9: Temporal Jittering Error (TJE). Lower values indi￾cate better temporal consistency. TJE measures the discrepancy between motion differences in real and generated videos, captur￾ing subtle jitter and flickering artifacts across frames. 4.7. Ethical Considerations and Po…
Figure 8
Figure 8. Figure 8: Cross-identity reenactment. Motions from a driving sub￾ject are transferred to a target avatar while preserving identity and gestures. 4.6.3. Without LPIPS loss. Finally, we remove the LPIPS term from the RGB reconstruc￾tion objective. This leads to lower perceptual qu…

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

184 extracted references · 37 canonical work pages

  1. [1]

    J. M. Buhmann and D. W. Fellner and M. Held and J. Ketterer and J. Puzicha , TITLE =. 1998 , PAGES =. doi:10.1111/1467-8659.00269 , NOTE =

    arXiv 1998

  2. [2]

    and Helmberg, Christoph , TITLE =

    Fellner, Dieter W. and Helmberg, Christoph , TITLE =. 1993 , PAGES =

    1993

  3. [3]

    Kobbelt and M

    L. Kobbelt and M. Stamminger and H.-P. Seidel , title =. doi:10.1111/1467-8659.16.3conferenceissue.36 , note =

    doi:10.1111/1467-8659.16.3conferenceissue.36

  4. [4]

    Lafortune and Sing-Choong Foo and Kenneth E

    Eric P. Lafortune and Sing-Choong Foo and Kenneth E. Torrance and Donald P. Greenberg , title =. Proc. SIGGRAPH '97 , volume = 31, pages =

  5. [5]

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

    Thin-plate spline motion model for image animation , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

  6. [6]

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

    One-shot free-view neural talking-head synthesis for video conferencing , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

  7. [7]

    Advances in Neural Information Processing Systems , volume=

    First order motion model for image animation , author=. Advances in Neural Information Processing Systems , volume=

  8. [8]

    Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition , month =

    Kossaifi, Jean and Tran, Linh and Panagakis, Yannis and Pantic, Maja , title =. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition , month =

Show all 184 references
  1. [9]

    Advances in Neural Information Processing Systems , volume=

    Improving 3d-aware image synthesis with a geometry-aware discriminator , author=. Advances in Neural Information Processing Systems , volume=

  2. [10]

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

    Pirenderer: Controllable portrait image generation via semantic neural rendering , author=. Proceedings of the IEEE/CVF International Conference on Computer Vision , pages=

  3. [11]

    Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition , month =

    Thies, Justus and Zollhofer, Michael and Stamminger, Marc and Theobalt, Christian and Niessner, Matthias , title =. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition , month =

  4. [12]

    Acm Transactions on Graphics (TOG) , volume=

    Deferred neural rendering: Image synthesis using neural textures , author=. Acm Transactions on Graphics (TOG) , volume=. 2019 , publisher=

    2019

  5. [13]

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

    Prinzler, Malte and Hilliges, Otmar and Thies, Justus , title =. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , month =. 2023 , pages =

    2023

  6. [14]

    Neural Head Avatars From Monocular RGB Videos , booktitle =

    Grassal, Philip-William and Prinzler, Malte and Leistner, Titus and Rother, Carsten and Nie. Neural Head Avatars From Monocular RGB Videos , booktitle =. 2022 , pages =

    2022

  7. [15]

    arXiv preprint arXiv:1807.11346 , year=

    Dropout-gan: Learning from a dynamic ensemble of discriminators , author=. arXiv preprint arXiv:1807.11346 , year=

    arXiv

  8. [16]

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

    Ensembling off-the-shelf models for gan training , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

  9. [17]

    Advances in Neural Information Processing Systems , volume=

    Dual discriminator generative adversarial nets , author=. Advances in Neural Information Processing Systems , volume=

  10. [18]

    and Hatef, M

    Kittler, J. and Hatef, M. and Duin, R.P.W. and Matas, J. , journal=. On combining classifiers , year=

  11. [19]

    Communications of the ACM , volume=

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

    2020

  12. [20]

    arXiv preprint arXiv:1511.06434 , year=

    Unsupervised representation learning with deep convolutional generative adversarial networks , author=. arXiv preprint arXiv:1511.06434 , year=

    arXiv

  13. [21]

    International conference on machine learning , pages=

    Wasserstein generative adversarial networks , author=. International conference on machine learning , pages=. 2017 , organization=

    2017

  14. [22]

    Advances in Neural Information Processing Systems , volume=

    Improved training of wasserstein gans , author=. Advances in Neural Information Processing Systems , volume=

  15. [23]

    Joon Son Chung and Arsha Nagrani and Andrew Zisserman , title=. Proc. Interspeech 2018 , pages=. doi:10.21437/Interspeech.2018-1929 , issn=

    2018 doi

  16. [24]

    Advances in Neural Information Processing Systems , volume=

    FNeVR: Neural volume rendering for face animation , author=. Advances in Neural Information Processing Systems , volume=

  17. [25]

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

    Depth-aware generative adversarial network for talking head video generation , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

  18. [26]

    Computer Vision--ECCV 2020: 16th European Conference, Glasgow, UK, August 23--28, 2020, Proceedings, Part XII 16 , pages=

    Fast bi-layer neural synthesis of one-shot realistic head avatars , author=. Computer Vision--ECCV 2020: 16th European Conference, Glasgow, UK, August 23--28, 2020, Proceedings, Part XII 16 , pages=. 2020 , organization=

    2020

  19. [27]

    Medical Image Computing and Computer-Assisted Intervention--MICCAI 2015: 18th International Conference, Munich, Germany, October 5-9, 2015, Proceedings, Part III 18 , pages=

    U-net: Convolutional networks for biomedical image segmentation , author=. Medical Image Computing and Computer-Assisted Intervention--MICCAI 2015: 18th International Conference, Munich, Germany, October 5-9, 2015, Proceedings, Part III 18 , pages=. 2015 , organization=

    2015

  20. [28]

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

    Unsupervised learning of probably symmetric deformable 3d objects from images in the wild , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

  21. [29]

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

    De-rendering 3D Objects in the Wild , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

  22. [30]

    arXiv preprint arXiv:2006.12057 , year=

    Differentiable rendering: A survey , author=. arXiv preprint arXiv:2006.12057 , year=

    2006 arXiv

  23. [31]

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

    Few-shot adversarial learning of realistic neural talking head models , author=. Proceedings of the IEEE/CVF international conference on computer vision , pages=

  24. [32]

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

    High-Fidelity and Freely Controllable Talking Head Video Generation , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

  25. [33]

    Advances in Neural Information Processing Systems , volume=

    Implicit warping for animation with image sets , author=. Advances in Neural Information Processing Systems , volume=

  26. [34]

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

    NeRFInvertor: High Fidelity NeRF-GAN Inversion for Single-shot Real Image Animation , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

  27. [35]

    IEEE signal processing letters , volume=

    Joint face detection and alignment using multitask cascaded convolutional networks , author=. IEEE signal processing letters , volume=. 2016 , publisher=

    2016

  28. [36]

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

    Low bandwidth video-chat compression using deep generative models , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

  29. [37]

    Proceedings of the 26th annual conference on Computer graphics and interactive techniques , pages=

    A morphable model for the synthesis of 3D faces , author=. Proceedings of the 26th annual conference on Computer graphics and interactive techniques , pages=

  30. [38]

    IEEE transactions on pattern analysis and machine intelligence , volume=

    Inverse rendering of faces with a 3D morphable model , author=. IEEE transactions on pattern analysis and machine intelligence , volume=. 2012 , publisher=

    2012

  31. [39]

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

    Inverserendernet: Learning single image inverse rendering , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

  32. [40]

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

    Image-to-image translation with conditional adversarial networks , author=. Proceedings of the IEEE conference on computer vision and pattern recognition , pages=

  33. [41]

    arXiv preprint arXiv:1411.1784 , year=

    Conditional generative adversarial nets , author=. arXiv preprint arXiv:1411.1784 , year=

    arXiv

  34. [42]

    IEEE Transactions on Visualization and Computer Graphics , volume=

    Optical models for direct volume rendering , author=. IEEE Transactions on Visualization and Computer Graphics , volume=. 1995 , publisher=

    1995

  35. [43]

    Communications of the ACM , volume=

    Nerf: Representing scenes as neural radiance fields for view synthesis , author=. Communications of the ACM , volume=. 2021 , publisher=

    2021

  36. [44]

    Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition , month =

    Wang, Ting-Chun and Liu, Ming-Yu and Zhu, Jun-Yan and Tao, Andrew and Kautz, Jan and Catanzaro, Bryan , title =. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition , month =

  37. [45]

    International Conference on Learning Representations , year=

    Spectral normalization for generative adversarial networks , author=. International Conference on Learning Representations , year=

  38. [46]

    arXiv preprint arXiv:1409.1556 , year=

    Very deep convolutional networks for large-scale image recognition , author=. arXiv preprint arXiv:1409.1556 , year=

    arXiv

  39. [47]

    2015 , eprint=

    ImageNet Large Scale Visual Recognition Challenge , author=. 2015 , eprint=

    2015

  40. [48]

    European Conference on Computer Vision , pages=

    Perceptual losses for real-time style transfer and super-resolution , author=. European Conference on Computer Vision , pages=. 2016 , publisher=

    2016

  41. [49]

    2017 , eprint=

    Adam: A Method for Stochastic Optimization , author=. 2017 , eprint=

    2017

  42. [50]

    Journal of Machine Learning Research , year =

    Michael Gashler , title =. Journal of Machine Learning Research , year =

  43. [51]

    2015 , eprint=

    Deep Learning Face Attributes in the Wild , author=. 2015 , eprint=

    2015

  44. [52]

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

    Closed-form factorization of latent semantics in gans , author=. Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages=

  45. [53]

    European conference on computer vision , pages=

    Styleheat: One-shot high-resolution editable talking face generation via pre-trained stylegan , author=. European conference on computer vision , pages=. 2022 , organization=

    2022

  46. [54]

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

    High-fidelity Facial Avatar Reconstruction from Monocular Video with Generative Priors , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

  47. [55]

    International Conference on Learning Representations , year=

    Latent Image Animator: Learning to Animate Images via Latent Space Navigation , author=. International Conference on Learning Representations , year=

  48. [56]

    2021 International Conference on 3D Vision , pages=

    SAFA: Structure Aware Face Animation , author=. 2021 International Conference on 3D Vision , pages=. 2021 , organization=

    2021

  49. [57]

    , author=

    Learning a model of facial shape and expression from 4D scans. , author=. ACM Trans. Graph. , volume=

  50. [58]

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

    Gaussian Head Avatar: Ultra High-fidelity Head Avatar via Dynamic Gaussians , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , year=

  51. [59]

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

    Analyzing and improving the image quality of stylegan , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

  52. [60]

    Advances in Neural Information Processing Systems , volume=

    Denoising diffusion probabilistic models , author=. Advances in Neural Information Processing Systems , volume=

  53. [61]

    arXiv preprint arXiv:2404.10667 , year=

    VASA-1: Lifelike Audio-Driven Talking Faces Generated in Real Time , author=. arXiv preprint arXiv:2404.10667 , year=

    arXiv

  54. [62]

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

    Difftalk: Crafting diffusion models for generalized audio-driven portraits animation , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

  55. [63]

    arXiv preprint arXiv:2402.17485 , year=

    EMO: Emote Portrait Alive-Generating Expressive Portrait Videos with Audio2Video Diffusion Model under Weak Conditions , author=. arXiv preprint arXiv:2402.17485 , year=

    arXiv

  56. [64]

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

    HyperReenact: one-shot reenactment via jointly learning to refine and retarget faces , author=. Proceedings of the IEEE/CVF International Conference on Computer Vision , pages=

  57. [65]

    2019 , eprint=

    Semantic Image Synthesis with Spatially-Adaptive Normalization , author=. 2019 , eprint=

    2019

  58. [66]

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

    Dpe: Disentanglement of pose and expression for general video portrait editing , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

  59. [67]

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

    Megaportraits: One-shot megapixel neural head avatars , author=. Proceedings of the 30th ACM International Conference on Multimedia , pages=

  60. [68]

    arXiv preprint arXiv:2311.16060 , year=

    DiffSLVA: Harnessing Diffusion Models for Sign Language Video Anonymization , author=. arXiv preprint arXiv:2311.16060 , year=

    arXiv

  61. [69]

    arXiv preprint arXiv:2308.16082 , year=

    SignDiff: Learning Diffusion Models for American Sign Language Production , author=. arXiv preprint arXiv:2308.16082 , year=

    arXiv

  62. [70]

    2024 IEEE International Conference on Artificial Intelligence and eXtended and Virtual Reality (AIxVR) , pages=

    Avatar quality: A study on presence and user preference , author=. 2024 IEEE International Conference on Artificial Intelligence and eXtended and Virtual Reality (AIxVR) , pages=. 2024 , organization=

    2024

  63. [71]

    , author=

    Real-time expression transfer for facial reenactment. , author=. ACM Trans. Graph. , volume=

  64. [72]

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

    Diner: Depth-aware image-based neural radiance fields , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

  65. [73]

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

    Instant volumetric head avatars , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

  66. [74]

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

    Soft rasterizer: A differentiable renderer for image-based 3d reasoning , author=. Proceedings of the IEEE/CVF International Conference on Computer Vision , pages=

  67. [75]

    Kerbl, Bernhard and Kopanas, Georgios and Leimk. 3. ACM Trans. Graph. , volume=

  68. [76]

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

    Gaussian head avatar: Ultra high-fidelity head avatar via dynamic gaussians , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

  69. [77]

    Advances in Neural Information Processing Systems , volume=

    Generative adversarial nets , author=. Advances in Neural Information Processing Systems , volume=

  70. [78]

    Advances in Neural Information Processing Systems , volume=

    Diffusion models beat gans on image synthesis , author=. Advances in Neural Information Processing Systems , volume=

  71. [79]

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

    High-resolution image synthesis with latent diffusion models , author=. Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages=

  72. [80]

    Advances in Neural Information Processing Systems , volume=

    Photorealistic text-to-image diffusion models with deep language understanding , author=. Advances in Neural Information Processing Systems , volume=

  73. [81]

    Advances in Neural Information Processing Systems , volume=

    Video diffusion models , author=. Advances in Neural Information Processing Systems , volume=

  74. [82]

    arXiv preprint arXiv:2406.01900 , year=

    Follow-Your-Emoji: Fine-Controllable and Expressive Freestyle Portrait Animation , author=. arXiv preprint arXiv:2406.01900 , year=

    arXiv

  75. [83]

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

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

  76. [84]

    Forty-first International Conference on Machine Learning , year=

    MagicPose: Realistic Human Poses and Facial Expressions Retargeting with Identity-aware Diffusion , author=. Forty-first International Conference on Machine Learning , year=

  77. [85]

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

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

  78. [86]

    1980 , publisher=

    Linear Systems , author=. 1980 , publisher=

    1980

  79. [87]

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

    Zip-nerf: Anti-aliased grid-based neural radiance fields , author=. Proceedings of the IEEE/CVF International Conference on Computer Vision , pages=

  80. [88]

    ACM Transactions on Graphics (TOG) , volume=

    Learning Images Across Scales Using Adversarial Training , author=. ACM Transactions on Graphics (TOG) , volume=. 2024 , publisher=

    2024

  81. [89]

    Advances in Neural Information Processing Systems , volume=

    Alias-free generative adversarial networks , author=. Advances in Neural Information Processing Systems , volume=

  82. [90]

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

    Motion representations for articulated animation , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

  83. [91]

    Li, Tianye and Bolkart, Timo and Black, Michael. J. and Li, Hao and Romero, Javier , journal =. Learning a model of facial shape and expression from. 2017 , pages =

    2017

  84. [92]

    , title =

    Loper, Matthew and Mahmood, Naureen and Romero, Javier and Pons-Moll, Gerard and Black, Michael J. , title =. ACM Trans. Graphics (Proc. SIGGRAPH Asia) , month = oct, number =

  85. [93]

    Advances in Neural Information Processing Systems , volume=

    Implicit neural representations with periodic activation functions , author=. Advances in Neural Information Processing Systems , volume=

  86. [94]

    Dynamic Neural Radiance Fields for Monocular 4D Facial Avatar Reconstruction , booktitle =

    Gafni, Guy and Thies, Justus and Zollh. Dynamic Neural Radiance Fields for Monocular 4D Facial Avatar Reconstruction , booktitle =. 2021 , pages =

    2021

  87. [95]

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

    Animatable gaussians: Learning pose-dependent gaussian maps for high-fidelity human avatar modeling , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

  88. [96]

    European Conference on Computer Vision , pages=

    Face2face : Real-time high-resolution one-shot face reenactment , author=. European Conference on Computer Vision , pages=. 2022 , organization=

    2022

  89. [97]

    arXiv preprint arXiv:2408.13049 , year=

    G3FA: Geometry-guided GAN for Face Animation , author=. arXiv preprint arXiv:2408.13049 , year=

    arXiv

  90. [98]

    arXiv preprint arXiv:2409.01876 , year=

    CyberHost: Taming Audio-driven Avatar Diffusion Model with Region Codebook Attention , author=. arXiv preprint arXiv:2409.01876 , year=

    arXiv

  91. [99]

    arXiv preprint arXiv:2409.02634 , year=

    Loopy: Taming Audio-Driven Portrait Avatar with Long-Term Motion Dependency , author=. arXiv preprint arXiv:2409.02634 , year=

    arXiv

  92. [100]

    arXiv preprint arXiv:2406.19680 , year=

    MimicMotion: High-Quality Human Motion Video Generation with Confidence-aware Pose Guidance , author=. arXiv preprint arXiv:2406.19680 , year=

    arXiv

  93. [101]

    arXiv preprint arXiv:2407.09012 , year=

    TCAN: Animating Human Images with Temporally Consistent Pose Guidance using Diffusion Models , author=. arXiv preprint arXiv:2407.09012 , year=

    arXiv

  94. [102]

    arXiv preprint arXiv:2403.14781 , year=

    Champ: Controllable and consistent human image animation with 3d parametric guidance , author=. arXiv preprint arXiv:2403.14781 , year=

    arXiv

  95. [103]

    Journal of Machine Learning Research , volume=

    Cascaded diffusion models for high fidelity image generation , author=. Journal of Machine Learning Research , volume=

  96. [104]

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

    Structure and content-guided video synthesis with diffusion models , author=. Proceedings of the IEEE/CVF International Conference on Computer Vision , pages=

  97. [105]

    International Conference on Learning Representations , year=

    AnimateDiff: Animate Your Personalized Text-to-Image Diffusion Models without Specific Tuning , author=. International Conference on Learning Representations , year=

  98. [106]

    arXiv preprint arXiv:2311.15127 , year=

    Stable video diffusion: Scaling latent video diffusion models to large datasets , author=. arXiv preprint arXiv:2311.15127 , year=

    arXiv

  99. [107]

    arXiv preprint arXiv:2401.03048 , year=

    Latte: Latent Diffusion Transformer for Video Generation , author=. arXiv preprint arXiv:2401.03048 , year=

    arXiv

  100. [108]

    International conference on machine learning , pages=

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

    2021

  101. [109]

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

    Adding conditional control to text-to-image diffusion models , author=. Proceedings of the IEEE/CVF International Conference on Computer Vision , pages=

  102. [110]

    Proceedings of the AAAI Conference on Artificial Intelligence , volume=

    T2i-adapter: Learning adapters to dig out more controllable ability for text-to-image diffusion models , author=. Proceedings of the AAAI Conference on Artificial Intelligence , volume=

  103. [111]

    arXiv preprint arXiv:2206.07696 , year=

    Diffusion models for video prediction and infilling , author=. arXiv preprint arXiv:2206.07696 , year=

    arXiv

  104. [112]

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

    Align your latents: High-resolution video synthesis with latent diffusion models , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

  105. [113]

    arXiv preprint arXiv:2401.12945 , year=

    Lumiere: A space-time diffusion model for video generation , author=. arXiv preprint arXiv:2401.12945 , year=

    arXiv

  106. [114]

    arXiv preprint arXiv:2305.13077 , year=

    Controlvideo: Training-free controllable text-to-video generation , author=. arXiv preprint arXiv:2305.13077 , year=

    arXiv

  107. [115]

    2023 IEEE/CVF International Conference on Computer Vision (ICCV) , pages=

    Dreampose: Fashion image-to-video synthesis via stable diffusion , author=. 2023 IEEE/CVF International Conference on Computer Vision (ICCV) , pages=. 2023 , organization=

    2023

  108. [116]

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

    Disco: Disentangled control for realistic human dance generation , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

  109. [117]

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

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

  110. [118]

    arXiv preprint arXiv:2010.02502 , year=

    Denoising diffusion implicit models , author=. arXiv preprint arXiv:2010.02502 , year=

    2010 arXiv

  111. [119]

    2nd International Conference on Learning Representations (ICLR) , year=

    Auto-Encoding Variational Bayes , author=. 2nd International Conference on Learning Representations (ICLR) , year=

  112. [120]

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

    The unreasonable effectiveness of deep features as a perceptual metric , author=. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition , pages=

  113. [121]

    Advances in Neural Information Processing Systems , volume =

    Ashish Vaswani and Noam Shazeer and Niki Parmar and Jakob Uszkoreit and Llion Jones and Aidan N Gomez and Łukasz Kaiser and Illia Polosukhin , title =. Advances in Neural Information Processing Systems , volume =

  114. [122]

    Proceedings of the Third Workshop on Computer Vision for AR/VR at IEEE Computer Vision and Pattern Recognition (CVPR) , year =

    Camillo Lugaresi and Jiuqiang Tang and Hadon Nash and Chris McClanahan and Esha Uboweja and Michael Hays and Fan Zhang and Chuo-Ling Chang and Ming Yong and Juhyun Lee and Wan-Teh Chang and Wei Hua and Manfred Georg and Matthias Grundmann , title =. Proceedings of the Third Wo...

  115. [123]

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

    Realtime multi-person 2d pose estimation using part affinity fields , author=. Proceedings of the IEEE conference on computer vision and pattern recognition , pages=

  116. [124]

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

    Controllable person image synthesis with pose-constrained latent diffusion , author=. Proceedings of the IEEE/CVF International Conference on Computer Vision , pages=

  117. [125]

    2002 , publisher=

    Nonlinear Systems , author=. 2002 , publisher=

    2002

  118. [126]

    2020 , booktitle =

    Xu, Kun and Li, Chongxuan and Zhu, Jun and Zhang, Bo , title =. 2020 , booktitle =

    2020

  119. [127]

    2021 , booktitle =

    Ho, Jonathan and Salimans, Tim , title =. 2021 , booktitle =

    2021

  120. [128]

    IEEE Transactions on Image Processing , volume=

    Image quality assessment: from error visibility to structural similarity , author=. IEEE Transactions on Image Processing , volume=. 2004 , publisher=

    2004

  121. [129]

    SSIM , author=

    Image quality metrics: PSNR vs. SSIM , author=. 2010 20th International Conference on Pattern Recognition , pages=. 2010 , organization=

    2010

  122. [130]

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

    Encoding in style: a stylegan encoder for image-to-image translation , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

  123. [131]

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

    Curricularface: adaptive curriculum learning loss for deep face recognition , author=. proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages=

  124. [132]

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

    Arcface: Additive angular margin loss for deep face recognition , author=. Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages=

  125. [133]

    International Conference on Learning Representations (ICLR) , year=

    Towards accurate generative models of video: A new metric & challenges , author=. International Conference on Learning Representations (ICLR) , year=

  126. [134]

    arXiv preprint arXiv:2209.14792 , year=

    Make-a-video: Text-to-video generation without text-video data , author=. arXiv preprint arXiv:2209.14792 , year=

    arXiv

  127. [135]

    Neural networks , volume=

    Approximation capabilities of multilayer feedforward networks , author=. Neural networks , volume=. 1991 , publisher=

    1991

  128. [136]

    1995 , publisher=

    Discrete-time control systems , author=. 1995 , publisher=

    1995

  129. [137]

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

    Diffusionavatars: Deferred diffusion for high-fidelity 3d head avatars , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

  130. [138]

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

    Gaussianavatars: Photorealistic head avatars with rigged 3d gaussians , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

  131. [139]

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

    Mip-nerf: A multiscale representation for anti-aliasing neural radiance fields , author=. Proceedings of the IEEE/CVF international conference on computer vision , pages=

  132. [140]

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

    Stylegan-v: A continuous video generator with the price, image quality and perks of stylegan2 , author=. Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages=

  133. [141]

    arXiv preprint arXiv:1806.05622 , year=

    Voxceleb2: Deep speaker recognition , author=. arXiv preprint arXiv:1806.05622 , year=

    arXiv

  134. [142]

    European Conference on Computer Vision , pages=

    Film: Frame interpolation for large motion , author=. European Conference on Computer Vision , pages=. 2022 , organization=

    2022

  135. [143]

    Avatar quality: A study on presence and user preference , year=

    Gasch, Cristina and Javanmardi, Alireza and Garcia-Palacios, Azucena and Pagani, Alain , booktitle=. Avatar quality: A study on presence and user preference , year=

  136. [144]

    Applied Sciences , VOLUME =

    Gasch, Cristina and Javanmardi, Alireza and Khan, Ameer and Garcia-Palacios, Azucena and Pagani, Alain , TITLE =. Applied Sciences , VOLUME =. 2025 , NUMBER =

    2025

  137. [145]

    and Kageki, Norri , journal=

    Mori, Masahiro and MacDorman, Karl F. and Kageki, Norri , journal=. The Uncanny Valley [From the Field] , year=

  138. [146]

    arXiv preprint arXiv:2503.10625 , year=

    LHM: Large Animatable Human Reconstruction Model from a Single Image in Seconds , author=. arXiv preprint arXiv:2503.10625 , year=

    arXiv

  139. [147]

    3D Gaussian Splatting for Real-Time Radiance Field Rendering , journal =

    Kerbl, Bernhard and Kopanas, Georgios and Leimk. 3D Gaussian Splatting for Real-Time Radiance Field Rendering , journal =. 2023 , url =

    2023

  140. [148]

    2020 , booktitle=

    NeRF: Representing Scenes as Neural Radiance Fields for View Synthesis , author=. 2020 , booktitle=

    2020

  141. [149]

    and Bouaziz, Sofien and Goldman, Dan B and Martin-Brualla, Ricardo and Seitz, Steven M

    Park, Keunhong and Sinha, Utkarsh and Hedman, Peter and Barron, Jonathan T. and Bouaziz, Sofien and Goldman, Dan B and Martin-Brualla, Ricardo and Seitz, Steven M. , title =. ACM Trans. Graph. , issue_date =. 2021 , articleno =

    2021

  142. [150]

    2023 IEEE 17th International Conference on Automatic Face and Gesture Recognition (FG) , pages=

    Flame-in-nerf: Neural control of radiance fields for free view face animation , author=. 2023 IEEE 17th International Conference on Automatic Face and Gesture Recognition (FG) , pages=. 2023 , organization=

    2023

  143. [151]

    Proceedings of the SIGGRAPH Asia 2025 Conference Papers , pages=

    Audio Driven Universal Gaussian Head Avatars , author=. Proceedings of the SIGGRAPH Asia 2025 Conference Papers , pages=

    2025

  144. [152]

    Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) , month =

    Pang, Haokai and Zhu, Heming and Kortylewski, Adam and Theobalt, Christian and Habermann, Marc , title =. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) , month =. 2024 , pages =

    2024

  145. [153]

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

    Scaffoldavatar: High-fidelity gaussian avatars with patch expressions , author=. Proceedings of the Special Interest Group on Computer Graphics and Interactive Techniques Conference Conference Papers , pages=

  146. [154]

    CVPR , month =

    Synthetic Prior for Few-Shot Drivable Head Avatar Inversion , author =. CVPR , month =

  147. [155]

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

    LAM: Large Avatar Model for One-shot Animatable Gaussian Head , author=. Proceedings of the Special Interest Group on Computer Graphics and Interactive Techniques Conference Conference Papers , pages=

  148. [156]

    Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV) , month =

    Zhang, Dongbin and Liu, Yunfei and Lin, Lijian and Zhu, Ye and Li, Yang and Qin, Minghan and Li, Yu and Wang, Haoqian , title =. Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV) , month =. 2025 , pages =

    2025

  149. [157]

    LIA: Latent Image Animator , year=

    Wang, Yaohui and Yang, Di and Bremond, Francois and Dantcheva, Antitza , journal=. LIA: Latent Image Animator , year=

  150. [158]

    International Conference on Machine Learning , year=

    MimicMotion: High-Quality Human Motion Video Generation with Confidence-aware Pose Guidance , author=. International Conference on Machine Learning , year=

  151. [159]

    Proceedings of the Computer Vision and Pattern Recognition Conference , pages=

    Stableanimator: High-quality identity-preserving human image animation , author=. Proceedings of the Computer Vision and Pattern Recognition Conference , pages=

  152. [160]

    , title =

    Taubner, Felix and Zhang, Ruihang and Tuli, Mathieu and Lindell, David B. , title =. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) , month =. 2025 , pages =

    2025

  153. [161]

    Avat3r: Large Animatable Gaussian Reconstruction Model for High-fidelity 3D Head Avatars , booktitle =

    Kirschstein, Tobias and Romero, Javier and Sevastopolsky, Artem and Nie. Avat3r: Large Animatable Gaussian Reconstruction Model for High-fidelity 3D Head Avatars , booktitle =. 2025 , pages =

    2025

  154. [162]

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

    Effective whole-body pose estimation with two-stages distillation , author=. Proceedings of the IEEE/CVF International Conference on Computer Vision , pages=

  155. [163]

    35th British Machine Vision Conference 2024,

    Alireza Javanmardi and Alain Pagani and Didier Stricker , title =. 35th British Machine Vision Conference 2024,

    2024

  156. [164]

    Pavlakos, Georgios and Choutas, Vasileios and Ghorbani, Nima and Bolkart, Timo and Osman, Ahmed A. A. and Tzionas, Dimitrios and Black, Michael J. , booktitle =. Expressive Body Capture:

  157. [165]

    CAAI Artificial Intelligence Research , volume =

    Bilateral Reference for High-Resolution Dichotomous Image Segmentation , author=. CAAI Artificial Intelligence Research , volume =

  158. [166]

    European Conference on Computer Vision , pages=

    Sapiens: Foundation for human vision models , author=. European Conference on Computer Vision , pages=. 2024 , organization=

    2024

  159. [167]

    European conference on computer vision , pages=

    Towards metrical reconstruction of human faces , author=. European conference on computer vision , pages=. 2022 , organization=

    2022

  160. [168]

    Reconstructing Hands in 3

    Pavlakos, Georgios and Shan, Dandan and Radosavovic, Ilija and Kanazawa, Angjoo and Fouhey, David and Malik, Jitendra , booktitle=. Reconstructing Hands in 3

  161. [169]

    ACM Transactions on Graphics, (Proc

    Embodied Hands: Modeling and Capturing Hands and Bodies Together , author =. ACM Transactions on Graphics, (Proc. SIGGRAPH Asia) , volume =. 2017 , month_numeric =

    2017

  162. [170]

    T-PAMI , year=

    Fast and Robust Multi-Person 3D Pose Estimation and Tracking from Multiple Views , author=. T-PAMI , year=

  163. [171]

    Advances in Neural Information Processing Systems , volume=

    Expressive gaussian human avatars from monocular rgb video , author=. Advances in Neural Information Processing Systems , volume=

  164. [172]

    Expressive Whole-Body

    Gyeongsik Moon and Takaaki Shiratori and Shunsuke Saito , booktitle=. Expressive Whole-Body

  165. [173]

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

    Gart: Gaussian articulated template models , author=. Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages=

  166. [174]

    , citeulike-article-id =

    Bradski, G. , citeulike-article-id =. Dr. Dobb's Journal of Software Tools , keywords =

  167. [175]

    Pattern Recognition , volume =

    U2-Net: Going Deeper with Nested U-Structure for Salient Object Detection , author =. Pattern Recognition , volume =

  168. [176]

    An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale ,author =

  169. [177]

    IEEE Transactions on Pattern Analysis and Machine Intelligence , title =

    Z. IEEE Transactions on Pattern Analysis and Machine Intelligence , title =

  170. [178]

    and Nocedal, Jorge , journal=

    Liu, Dong C. and Nocedal, Jorge , journal=. On the limited memory. 1989 , publisher=

    1989

  171. [179]

    CVPR , year =

    Shunsuke Saito and Gabriel Schwartz and Tomas Simon and Junxuan Li and Giljoo Nam , title =. CVPR , year =

  172. [180]

    , author =

    EVA: Expressive Virtual Avatars from Multi-view Videos. , author =. SIGGRAPH 2025 Conference Papers , pages=

    2025

  173. [181]

    2024 , booktitle=

    Cafca: High-quality Novel View Synthesis of Expressive Faces from Casual Few-shot Captures , author=. 2024 , booktitle=

    2024

  174. [182]

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

    Facecraft4d: Animated 3d facial avatar generation from a single image , author=. Proceedings of the IEEE/CVF International Conference on Computer Vision , pages=

  175. [183]

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

    TalkingPose: Efficient Face and Gesture Animation with Feedback-guided Diffusion Model , author=. Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision , pages=

  176. [184]

    , journal=

    Bukimi no tani [The uncanny valley]. , journal=

Pith tools

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