Pith. sign in

REVIEW 3 major objections 5 minor 2 cited by

GaussianAvatar-Editor: Photorealistic Animatable Gaussian Head Avatar Editor

T0 review · 3 major / 5 minor · reviewed 2026-08-10 · deepseek-v4-flash

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

Pith's one-line read A weighted alpha blending equation makes text-driven editing of animatable Gaussian head avatars robust to motion occlusion, while adversarial learning keeps results consistent in time.

desk verdict WABE is a genuinely useful occlusion-aware editing trick; Eq. 7's asymmetric fake pairs undermine the temporal-consistency claim, likely a typo but needs fixing. read the letter →

arxiv 2501.09978 v1 pith:ZF3ACQEG submitted 2025-01-17 cs.CV

Xiangyue Liu , Kunming Luo , Heng Li , Qi Zhang , Yuan Liu , Li Yi , Ping Tan This is my paper · ORCID
classification cs.CV
keywords Gaussiansplattingheadavatareditingtext-drivenanimatablemotionocclusiontemporalconsistencyadversariallearningFLAMEmodel
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 introduces GaussianAvatar-Editor, a framework for editing animatable Gaussian head avatars from a text prompt while preserving the ability to reanimate with new poses, expressions, and viewpoints. It argues that standard 3D Gaussian Splatting alpha blending corrupts occluded parts such as teeth, eyeballs, and nostrils during editing, because gradients from visible pixels reach invisible Gaussians. To fix this, it replaces the blending weight with WABE, which assigns near-zero weight to occluded Gaussians. To fix spatial-temporal inconsistency in diffusion-edited supervision, it adds a conditional adversarial loss that compares rendered and edited image pairs across adjacent timesteps. The claimed result is photorealistic, consistent editing of 4D avatars with clean occluded regions.

What carries the argument

The central object is the Weighted Alpha Blending Equation (WABE), Eq. 4 and Eq. 5 of the paper: $C(x)=\sum_k w_k c_k \alpha_k \prod_{j<k}(1-\alpha_j)$ with $w_k = e^{-\beta(1-\prod_{j<k}(1-\alpha_j))}$. The factor $\prod_{j<k}(1-\alpha_j)$ is the accumulated visibility of all Gaussians in front of $k$; WABE multiplies each Gaussian's contribution by a weight that is approximately one when fully visible and near zero when occluded, so that occluded Gaussians (teeth behind lips, eyeballs behind eyelids) receive almost no update from image losses. The second mechanism is the temporal adversarial loss of Eq. 7, which feeds pairs $(E_i^t, E_i^t - E_i^k)$ as real and $(C_i^t, C_i^t - E_i^k)$ as fake to a discriminator, intended to make rendered frames temporally consistent with edited frames.

What would settle it

Train the discriminator of Eq. 7 with the pairing exactly as written and separately evaluate its accuracy on the first channel and on the second channel; if the first channel alone reaches near-perfect accuracy, the temporal-consistency signal attributed to the adversarial loss is not being delivered by the stated formula.

Watch

Extended reading notes

Core claim

The paper claims that text-driven editing of an animatable FLAME-bound Gaussian head avatar can be made robust to motion occlusion and 4D inconsistency by two changes to the render-edit-aggregate pipeline. First, replacing the 3DGS $\alpha$ blending equation with a weighted $\alpha$ blending equation (WABE), where each Gaussian's weight is $w_k = e^{-\beta(1 - \prod_{j<k}(1-\alpha_j))}$, suppresses occluded Gaussians so that gradients from visible pixels do not update them. Second, a conditional adversarial loss over image pairs from adjacent timesteps teaches the avatar temporal consistency beyond what the reconstruction loss provides. With these components, the edited avatar can be reenacted by the same subject or a different actor and rendered from novel views with consistent appearance.

Load-bearing premise

The temporal adversarial loss is assumed to teach consistency across timesteps, but as written the fake pair's second channel is $C_i^t - E_i^k$ while the real pair's second channel is $E_i^t - E_i^k$, so the discriminator can likely tell them apart from the first channel alone.

Editorial extensions

If this is right

  • Editing an avatar once with a text prompt yields a re-animatable asset: the edited appearance persists under new poses, expressions, and viewpoints without per-frame editing.
  • Occluded regions such as teeth, eyeballs, and nostrils remain unchanged by edits meant for visible skin, so prompts like 'add a beard' do not paint over teeth.
  • The same pipeline works on avatars reconstructed from multi-view video, with each edit taking about 15 minutes on a single A100.
  • Quantitative CLIP-S and CLIP-C scores improve over applying Instruct-NeRF2NeRF to GaussianAvatars or INSTA, and over the Control4D reimplementation, across novel-view, self-reenactment, and cross-identity settings.
  • The temporal adversarial loss is claimed to reduce blur and artifacts in animation by enforcing consistency across time steps.

Reading between the lines

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

  • WABE is not limited to heads: any multi-layer surface with self-occlusion, such as hands, bodies, or clothed avatars, could benefit from visibility-weighted blending during editing, provided the occupancy ordering is reliable.
  • If Eq. 7's fake pair is a typo, the correct pairing $(C_i^t, C_i^t - C_i^k)$ should be tested; if it is not a typo, the temporal-consistency contribution of the adversarial loss needs re-examination, since the discriminator can solve the classification from the first channel alone.
  • Because WABE suppresses occluded Gaussians, it may slow or block intended edits to surfaces that are frequently occluded, such as the inner mouth; users may need to expose those regions during optimization to edit them.
  • The paper's clip-based metrics measure text alignment, not temporal consistency directly; a dedicated temporal metric (for example, frame-to-frame correspondence or patch consistency) would sharpen the claim of 4D consistency.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 5 minor

Summary. GaussianAvatar-Editor proposes a text-driven editing framework for FLAME-bound animatable 3D Gaussian head avatars. The method follows a render-edit-aggregate loop based on Instruct-NeRF2NeRF, using InstructPix2Pix as the 2D editor. Two technical contributions are introduced: (i) the Weighted Alpha Blending Equation (WABE), which multiplies each Gaussian's alpha contribution by a visibility-dependent weight to suppress gradients from visible pixels toward occluded Gaussians, and (ii) a conditional adversarial loss intended to enforce spatial-temporal consistency during editing. The avatar is optimized on the NeRSemble dataset with reconstruction, adversarial, and regularization losses, and is evaluated through novel-view rendering, self-reenactment, and cross-identity reenactment using CLIP-S and CLIP-C metrics. The paper claims state-of-the-art editing quality and spatio-temporal consistency for animatable Gaussian avatars.

Significance. If the claims hold, this is a useful advance: it is among the first text-driven editing methods specifically designed for animatable Gaussian head avatars, and the occlusion problem it targets is real and visible in the qualitative comparisons. The paper's main strengths are the clear problem formulation, the qualitative WABE ablation in Figs. 3 and 10 (which convincingly shows that teeth, eyeballs, and nostrils are protected when WABE is enabled), and the practical render-edit-aggregate pipeline that achieves editing in about 15 minutes. The authors also commit to releasing code, which would help reproducibility. The quantitative evidence is less strong, however: Table 1 reports single runs without error bars, the Control4D baseline is a reimplementation, and the temporal-consistency mechanism as written in Eq. (7) is questionable. These issues affect the load-bearing claims and require clarification or additional experiments.

major comments (3)
  1. [Sec. 4.3, Eq. (7)]
  2. [Sec. 4.2 vs. Sec. 4.5]
  3. [Table 1 and Sec. 5.1]
minor comments (5)
  1. [Sec. 4.3, Eq. (6)]
  2. [Table 1]
  3. [Sec. 5.1, Implementation details]
  4. [Sec. 5.1, Dataset]
  5. [Sec. 4.2, Eq. (5)]

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the WABE visibility weighting and the adversarial losses are self-contained design choices validated by ablation, with only a motivational self-citation that is not load-bearing.

full rationale

The paper's central claimed derivation chain is not circular. WABE (Eq. 4 and Eq. 5) is introduced as a modified alpha-blending rule in which each Gaussian's weight is a monotone function of its accumulated visibility; the claim that occluded Gaussians receive small weight follows directly from the definition, but the paper does not present this as a prediction derived from independent inputs. Its effectiveness is tested by disabling WABE (Table 1, Fig. 10), which is an ablation, not a fitted parameter renamed as a result. The adversarial losses in Eq. 7 are defined in the paper and optimized jointly with the reconstruction loss; the temporal-consistency claim is supported by an explicit ablation ('Ours w/o adv'). Reference [17] (GenN2N) is a self-citation by overlapping authors, but it is used only as motivation for conditional adversarial training, not as a load-bearing proof or a uniqueness theorem, so it does not make the derivation circular. The manuscript's own limitation statement (Sec. 6) about FLAME not animating the tongue is an acknowledged modeling limitation, not a hidden circular dependency. One non-circular correctness concern should be noted but does not affect the circularity score: in Eq. 7 the real pair is (E_i^t, E_i^t - E_i^k) and the fake pair is (C_i^t, C_i^t - E_i^k), so the first channel already differs between real and fake, and the second channel mixes a rendered frame with an edited frame from an adjacent timestep; this may undermine the temporal-consistency interpretation of the loss, but it is a potential implementation flaw rather than an equivalence of the result to its inputs. The evaluation uses standard external CLIP metrics and compares against external baselines, so the central method is not self-validating by construction.

Assumptions & free parameters 2 free parameters · 5 assumptions · 0 invented entities

The central method rests on prior reconstruction and image-editing components (GaussianAvatars, InstructPix2Pix) plus two hand-designed additions: the WABE weighting and the temporal adversarial loss. Neither addition is derived, and both depend on manually chosen hyperparameters.

free parameters (2)
  • β (WABE sharpness) = 6
    Hand-set in Eq. 5 for all experiments; controls how rapidly visible Gaussians are favored over occluded ones.
  • Loss weights λ1, λ2, λ3, λ4 = 10, 0.01, 0.01, 10
    Hand-set weights in Eq. 8 balancing reconstruction, discriminator, generator adversarial, and Gaussian-avatar constraint losses.
assumptions (5)
  • standard math 3D Gaussian splatting alpha blending (Eq. 2) and its differentiable rendering
    The WABE modification in Eq. 4 builds directly on the standard 3DGS compositing model from [14].
  • domain assumption A FLAME-bound GaussianAvatar reconstruction exists for each subject
    The editing pipeline starts from a GaussianAvatars reconstruction [25]; editing quality inherits its reconstruction quality.
  • domain assumption InstructPix2Pix provides editing supervision aligned with user text for head images
    The render-edit-aggregate loop relies on [3] to generate target edited images E_t_i from the current render.
  • ad hoc to paper WABE's exponential form suppresses gradients to occluded Gaussians without harming visible regions
    The specific form w_k = exp(-β(1-T_k)) in Eq. 5 is asserted with β=6; no derivation or formal analysis is provided.
  • ad hoc to paper The pair construction in Eq. 7 is a valid conditional adversarial objective for temporal consistency
    The paper claims temporal consistency from this loss, but the fake pair mixes rendered and edited images asymmetrically and the mechanism is not demonstrated.

how reviews work

0 comments
Cite this review

Pith. "Pith review of GaussianAvatar-Editor: Photorealistic Animatable Gaussian Head Avatar Editor." pith.science (2026). https://pith.science/paper/ZF3ACQEG

@misc{pith2026250109978,
  author       = {Pith},
  title        = {Pith review of: GaussianAvatar-Editor: Photorealistic Animatable Gaussian Head Avatar Editor},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/ZF3ACQEG}},
  note         = {Machine review of arXiv:2501.09978}
}
read the original abstract

We introduce GaussianAvatar-Editor, an innovative framework for text-driven editing of animatable Gaussian head avatars that can be fully controlled in expression, pose, and viewpoint. Unlike static 3D Gaussian editing, editing animatable 4D Gaussian avatars presents challenges related to motion occlusion and spatial-temporal inconsistency. To address these issues, we propose the Weighted Alpha Blending Equation (WABE). This function enhances the blending weight of visible Gaussians while suppressing the influence on non-visible Gaussians, effectively handling motion occlusion during editing. Furthermore, to improve editing quality and ensure 4D consistency, we incorporate conditional adversarial learning into the editing process. This strategy helps to refine the edited results and maintain consistency throughout the animation. By integrating these methods, our GaussianAvatar-Editor achieves photorealistic and consistent results in animatable 4D Gaussian editing. We conduct comprehensive experiments across various subjects to validate the effectiveness of our proposed techniques, which demonstrates the superiority of our approach over existing methods. More results and code are available at: [Project Link](https://xiangyueliu.github.io/GaussianAvatar-Editor/).

Figures

Figures reproduced from arXiv: 2501.09978 by the authors.

Figure 1
Figure 1. We introduce GaussianAvatar-Editor, a method for text-driven editing of animatable Gaussian head avatars with fully controllable [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. The overview of our method. We follow a render-edit-aggregate optimization pipeline as in Instruct-NeRF2NeRF [ [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Illustration of the Weighted alpha blending equation [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figures from the paper (7 more)
Figure 4
Figure 4. Figure 4: Our results on novel view synthesis. We show our edited results using the text prompt [PITH_FULL_IMAGE:figures/full_fig_p005_4.png]
Figure 5
Figure 5. Figure 5: Comparison on novel view synthesis. Our method produces more high-quality and multi-view consistent results than baselines. [PITH_FULL_IMAGE:figures/full_fig_p005_5.png]
Figure 6
Figure 6. Figure 6: Our results on self-reenactment. Self-reenactment renders held-out unseen head pose and expressions from 16 training camera [PITH_FULL_IMAGE:figures/full_fig_p006_6.png]
Figure 7
Figure 7. Figure 7: Comparison of self-reenactment. Our edited avatar can correctly produce detailed facial features under unseen expressions and [PITH_FULL_IMAGE:figures/full_fig_p006_7.png]
Figure 8
Figure 8. Figure 8: Our results on cross-identity reenactment. Cross-identity reenactment animates the avatar to render images with unseen head [PITH_FULL_IMAGE:figures/full_fig_p007_8.png]
Figure 9
Figure 9. Figure 9: Comparison of Cross-identity reenactment. Different edited avatars are controlled by the same source actor. Our method can [PITH_FULL_IMAGE:figures/full_fig_p007_9.png]
Figure 10
Figure 10. Figure 10: Ablation study of WABE. Novel view rendering Self-reenactment Cross-identity reenactment CLIP-S ↑ CLIP-C ↑ CLIP-S ↑ CLIP-C ↑ CLIP-S ↑ CLIP-C ↑ INSTA+I-N2N 0.181 0.955 0.042 0.923 0.043 0.936 GA+I-N2N 0.236 0.968 0.044 0.938 0.069 0.941 Control4D 0.222 0.980 0.058 0.93…

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 2 Pith papers

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

  1. SVG-Head: Hybrid Surface-Volumetric Gaussians for High-Fidelity Head Reconstruction and Real-Time Editing

    cs.CV 2025-08 conditional novelty 6.0 of 10

    A hybrid of surface and volumetric 3D Gaussians tied to a FLAME mesh, with surface Gaussians sampling from explicit texture images, yields high-fidelity head avatars with real-time texture editing.

  2. Generative AI for Character Animation: A Comprehensive Survey of Techniques, Applications, and Future Directions

    cs.CV 2025-04 conditional novelty 3.0 of 10

    A comprehensive survey that unifies generative AI techniques for character animation across facial, gesture, motion, and 3D asset generation, with a shared taxonomy and resource list.

Reference graph

Works this paper leans on

45 extracted references · 26 canonical work pages · cited by 2 Pith papers

  1. [1]

    Sine: Semantic-driven image-based nerf editing with prior-guided editing field

    Chong Bao, Yinda Zhang, Bangbang Yang, Tianxing Fan, Zesong Yang, Hujun Bao, Guofeng Zhang, and Zhaopeng Cui. Sine: Semantic-driven image-based nerf editing with prior-guided editing field. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages 20919–20929, 2023. 2

    work page 2023

  2. [2]

    Text2live: Text-driven layered image and video editing

    Omer Bar-Tal, Dolev Ofri-Amar, Rafail Fridman, Yoni Kas- ten, and Tali Dekel. Text2live: Text-driven layered image and video editing. In European conference on computer vi- sion, pages 707–723. Springer, 2022. 2

    2022

  3. [3]

    In- structpix2pix: Learning to follow image editing instructions

    Tim Brooks, Aleksander Holynski, and Alexei A Efros. In- structpix2pix: Learning to follow image editing instructions. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 18392–18402, 2023. 2, 4, 5

    work page 2023

  4. [4]

    Effi- cient geometry-aware 3d generative adversarial networks

    Eric R Chan, Connor Z Lin, Matthew A Chan, Koki Nagano, Boxiao Pan, Shalini De Mello, Orazio Gallo, Leonidas J Guibas, Jonathan Tremblay, Sameh Khamis, et al. Effi- cient geometry-aware 3d generative adversarial networks. In CVPR, pages 16123–16133, 2022. 2

    2022

  5. [5]

    Monogaus- sianavatar: Monocular gaussian point-based head avatar

    Yufan Chen, Lizhen Wang, Qijing Li, Hongjiang Xiao, Shengping Zhang, Hongxun Yao, and Yebin Liu. Monogaus- sianavatar: Monocular gaussian point-based head avatar. arXiv, 2023. 2, 3

    work page 2023

  6. [6]

    Headgas: Real-time animatable head avatars via 3d gaussian splatting

    Helisa Dhamo, Yinyu Nie, Arthur Moreau, Jifei Song, Richard Shaw, Yiren Zhou, and Eduardo P ´erez-Pellitero. Headgas: Real-time animatable head avatars via 3d gaussian splatting. arXiv preprint arXiv:2312.02902, 2023. 2

    arXiv 2023

  7. [7]

    Dynamic neural radiance fields for monocular 4d facial avatar reconstruction

    Guy Gafni, Justus Thies, Michael Zollhofer, and Matthias Nießner. Dynamic neural radiance fields for monocular 4d facial avatar reconstruction. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages 8649–8658, 2021. 2

    work page 2021

  8. [8]

    Textdeformer: Geometry manipu- lation using text guidance

    William Gao, Noam Aigerman, Thibault Groueix, V ova Kim, and Rana Hanocka. Textdeformer: Geometry manipu- lation using text guidance. In ACM SIGGRAPH 2023 Con- ference Proceedings, pages 1–11, 2023. 2

    2023

Show all 45 references
  1. [9]

    Instruct-nerf2nerf: Edit- ing 3d scenes with instructions

    Ayaan Haque, Matthew Tancik, Alexei A Efros, Aleksander Holynski, and Angjoo Kanazawa. Instruct-nerf2nerf: Edit- ing 3d scenes with instructions. In Proceedings of the IEEE/CVF International Conference on Computer Vision , pages 19740–19750, 2023. 2, 3, 4, 5, 7

    2023

  2. [10]

    Prompt-to-prompt im- age editing with cross attention control

    Amir Hertz, Ron Mokady, Jay Tenenbaum, Kfir Aberman, Yael Pritch, and Daniel Cohen-Or. Prompt-to-prompt im- age editing with cross attention control. arXiv preprint arXiv:2208.01626, 2022. 2

    2022 arXiv

  3. [11]

    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. 5

    2016

  4. [12]

    Instruct 3d-to-3d: Text in- struction guided 3d-to-3d conversion

    Hiromichi Kamata, Yuiko Sakuma, Akio Hayakawa, Masato Ishii, and Takuya Narihira. Instruct 3d-to-3d: Text in- struction guided 3d-to-3d conversion. arXiv preprint arXiv:2303.15780, 2023. 2

    2023 arXiv

  5. [13]

    Imagic: Text-based real image editing with diffusion models

    Bahjat Kawar, Shiran Zada, Oran Lang, Omer Tov, Huiwen Chang, Tali Dekel, Inbar Mosseri, and Michal Irani. Imagic: Text-based real image editing with diffusion models. InPro- ceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 6007–6017, 2023. 2

    2023

  6. [14]

    3d gaussian splatting for real-time radiance field rendering

    Bernhard Kerbl, Georgios Kopanas, Thomas Leimk ¨uhler, and George Drettakis. 3d gaussian splatting for real-time radiance field rendering. ACM Transactions on Graphics (ToG), 42(4):1–14, 2023. 1, 3, 4

    2023

  7. [15]

    Nersemble: Multi-view radi- ance field reconstruction of human heads.ACM Transactions on Graphics (TOG), 42(4):1–14, 2023

    Tobias Kirschstein, Shenhan Qian, Simon Giebenhain, Tim Walter, and Matthias Nießner. Nersemble: Multi-view radi- ance field reconstruction of human heads.ACM Transactions on Graphics (TOG), 42(4):1–14, 2023. 6

    2023

  8. [16]

    Learning a model of facial shape and expression from 4d scans

    Tianye Li, Timo Bolkart, Michael J Black, Hao Li, and Javier Romero. Learning a model of facial shape and expression from 4d scans. ACM Trans. Graph., 36(6):194–1, 2017. 3, 6

    2017

  9. [17]

    Genn2n: Generative nerf2nerf translation

    Xiangyue Liu, Han Xue, Kunming Luo, Ping Tan, and Li Yi. Genn2n: Generative nerf2nerf translation. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 5105–5114, 2024. 5

    2024

  10. [18]

    Sdedit: Guided image synthesis and editing with stochastic differential equa- tions

    Chenlin Meng, Yutong He, Yang Song, Jiaming Song, Jia- jun Wu, Jun-Yan Zhu, and Stefano Ermon. Sdedit: Guided image synthesis and editing with stochastic differential equa- tions. arXiv preprint arXiv:2108.01073, 2021. 2

    2021 arXiv

  11. [19]

    Nerf: Representing scenes as neural radiance fields for view synthesis

    B Mildenhall, PP Srinivasan, M Tancik, JT Barron, R Ra- mamoorthi, and R Ng. Nerf: Representing scenes as neural radiance fields for view synthesis. In European conference on computer vision, 2020. 1

    2020

  12. [20]

    Nerf: Representing scenes as neural radiance fields for view syn- thesis

    Ben Mildenhall, Pratul P Srinivasan, Matthew Tancik, Jonathan T Barron, Ravi Ramamoorthi, and Ren Ng. Nerf: Representing scenes as neural radiance fields for view syn- thesis. Communications of the ACM , 65(1):99–106, 2021. 7

    2021

  13. [21]

    Null-text inversion for editing real im- ages using guided diffusion models

    Ron Mokady, Amir Hertz, Kfir Aberman, Yael Pritch, and Daniel Cohen-Or. Null-text inversion for editing real im- ages using guided diffusion models. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 6038–6047, 2023. 2

    2023

  14. [22]

    Avatarstudio: Text-driven edit- ing of 3d dynamic human head avatars

    Mohit Mendiratta Pan, Mohamed Elgharib, Kartik Teo- tia, Ayush Tewari, Vladislav Golyanik, Adam Kortylewski, Christian Theobalt, et al. Avatarstudio: Text-driven edit- ing of 3d dynamic human head avatars. arXiv preprint arXiv:2306.00547, 2023. 3

    2023 arXiv

  15. [23]

    Styleavatar: Stylizing animatable head avatars

    Juan C P ´erez, Thu Nguyen-Phuoc, Chen Cao, Art- siom Sanakoyeu, Tomas Simon, Pablo Arbel ´aez, Bernard Ghanem, Ali Thabet, and Albert Pumarola. Styleavatar: Stylizing animatable head avatars. In Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision,...

    2024

  16. [24]

    Dreamfusion: Text-to-3d using 2d diffusion

    Ben Poole, Ajay Jain, Jonathan T Barron, and Ben Milden- hall. Dreamfusion: Text-to-3d using 2d diffusion. arXiv preprint arXiv:2209.14988, 2022. 2

    2022 arXiv

  17. [25]

    Gaus- sianavatars: Photorealistic head avatars with rigged 3d gaus- sians

    Shenhan Qian, Tobias Kirschstein, Liam Schoneveld, Davide Davoli, Simon Giebenhain, and Matthias Nießner. Gaus- sianavatars: Photorealistic head avatars with rigged 3d gaus- sians. arXiv preprint arXiv:2312.02069, 2023. 2, 3, 4, 6, 7, 8 9

    2023 arXiv

  18. [26]

    Texture: Text-guided texturing of 3d shapes

    Elad Richardson, Gal Metzer, Yuval Alaluf, Raja Giryes, and Daniel Cohen-Or. Texture: Text-guided texturing of 3d shapes. arXiv preprint arXiv:2302.01721, 2023. 2

    2023 arXiv

  19. [27]

    High-resolution image synthesis with latent diffusion models

    Robin Rombach, Andreas Blattmann, Dominik Lorenz, Patrick Esser, and Bj ¨orn Ommer. High-resolution image synthesis with latent diffusion models. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 10684–10695, 2022. 2

    2022

  20. [28]

    Photorealistic text-to-image diffusion models with deep language understanding

    Chitwan Saharia, William Chan, Saurabh Saxena, Lala Li, Jay Whang, Emily L Denton, Kamyar Ghasemipour, Raphael Gontijo Lopes, Burcu Karagol Ayan, Tim Salimans, et al. Photorealistic text-to-image diffusion models with deep language understanding. Advances in Neural Information...

    2022

  21. [29]

    Relightable gaussian codec avatars

    Shunsuke Saito, Gabriel Schwartz, Tomas Simon, Junxuan Li, and Giljoo Nam. Relightable gaussian codec avatars. arXiv preprint arXiv:2312.03704, 2023. 2

    2023 arXiv

  22. [30]

    Relightable gaussian codec avatars

    Shunsuke Saito, Gabriel Schwartz, Tomas Simon, Junxuan Li, and Giljoo Nam. Relightable gaussian codec avatars. In CVPR, 2024. 2, 3

    2024

  23. [31]

    V ox-e: Text-guided voxel editing of 3d ob- jects

    Etai Sella, Gal Fiebelman, Peter Hedman, and Hadar Averbuch-Elor. V ox-e: Text-guided voxel editing of 3d ob- jects. In Proceedings of the IEEE/CVF International Con- ference on Computer Vision, pages 430–440, 2023. 2

    2023

  24. [32]

    Con- trol4d: Dynamic portrait editing by learning 4d gan from 2d diffusion-based editor

    Ruizhi Shao, Jingxiang Sun, Cheng Peng, Zerong Zheng, Boyao Zhou, Hongwen Zhang, and Yebin Liu. Con- trol4d: Dynamic portrait editing by learning 4d gan from 2d diffusion-based editor. arXiv preprint arXiv:2305.20082,

    arXiv

  25. [33]

    Blending-nerf: Text-driven localized editing in neural radiance fields

    Hyeonseop Song, Seokhun Choi, Hoseok Do, Chul Lee, and Taehyeong Kim. Blending-nerf: Text-driven localized editing in neural radiance fields. In Proceedings of the IEEE/CVF International Conference on Computer Vision , pages 14383–14393, 2023. 2

    2023

  26. [34]

    Sketch-guided text-to-image diffusion models

    Andrey V oynov, Kfir Aberman, and Daniel Cohen-Or. Sketch-guided text-to-image diffusion models. In ACM SIG- GRAPH 2023 Conference Proceedings , pages 1–11, 2023. 2

    2023

  27. [35]

    Clip-nerf: Text-and-image driven manip- ulation of neural radiance fields

    Can Wang, Menglei Chai, Mingming He, Dongdong Chen, and Jing Liao. Clip-nerf: Text-and-image driven manip- ulation of neural radiance fields. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 3835–3844, 2022. 2

    2022

  28. [36]

    Nerf-art: Text-driven neural radiance fields stylization

    Can Wang, Ruixiang Jiang, Menglei Chai, Mingming He, Dongdong Chen, and Jing Liao. Nerf-art: Text-driven neural radiance fields stylization. IEEE Transactions on Visualiza- tion and Computer Graphics, 2023. 2

    2023

  29. [37]

    Gaussianhead: High-fidelity head avatars with learnable gaussian derivation, 2024

    Jie Wang, Jiu-Cheng Xie, Xianyan Li, Feng Xu, Chi-Man Pun, and Hao Gao. Gaussianhead: High-fidelity head avatars with learnable gaussian derivation, 2024. 2, 3

    2024

  30. [38]

    Stylea- vatar: Real-time photo-realistic portrait avatar from a single video

    Lizhen Wang, Xiaochen Zhao, Jingxiang Sun, Yuxiang Zhang, Hongwen Zhang, Tao Yu, and Yebin Liu. Stylea- vatar: Real-time photo-realistic portrait avatar from a single video. arXiv preprint arXiv:2305.00942, 2023. 2

    2023 arXiv

  31. [39]

    Gaussian head avatar: Ultra high-fidelity head avatar via dynamic gaussians

    Yuelang Xu, Benwang Chen, Zhe Li, Hongwen Zhang, Lizhen Wang, Zerong Zheng, and Yebin Liu. Gaussian head avatar: Ultra high-fidelity head avatar via dynamic gaussians. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2024. 2, 3

    2024

  32. [40]

    Adding conditional control to text-to-image diffusion models

    Lvmin Zhang, Anyi Rao, and Maneesh Agrawala. Adding conditional control to text-to-image diffusion models. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 3836–3847, 2023. 2

    2023

  33. [41]

    Psavatar: A point-based morphable shape model for real-time head avatar creation with 3d gaussian splatting

    Zhongyuan Zhao, Zhenyu Bao, Qing Li, Guoping Qiu, and Kanglin Liu. Psavatar: A point-based morphable shape model for real-time head avatar creation with 3d gaussian splatting. arXiv preprint arXiv:2401.12900, 2024. 2

    2024 arXiv

  34. [42]

    Im avatar: Implicit morphable head avatars from videos

    Yufeng Zheng, Victoria Fern ´andez Abrevaya, Marcel C B¨uhler, Xu Chen, Michael J Black, and Otmar Hilliges. Im avatar: Implicit morphable head avatars from videos. InPro- ceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 13545–13555, 2022. 2

    2022

  35. [43]

    Dreameditor: Text-driven 3d scene editing with neural fields

    Jingyu Zhuang, Chen Wang, Liang Lin, Lingjie Liu, and Guanbin Li. Dreameditor: Text-driven 3d scene editing with neural fields. In SIGGRAPH Asia 2023 Conference Papers, pages 1–10, 2023. 2

    2023

  36. [44]

    Tip-editor: An accurate 3d editor fol- lowing both text-prompts and image-prompts.arXiv preprint arXiv:2401.14828, 2024

    Jingyu Zhuang, Di Kang, Yan-Pei Cao, Guanbin Li, Liang Lin, and Ying Shan. Tip-editor: An accurate 3d editor fol- lowing both text-prompts and image-prompts.arXiv preprint arXiv:2401.14828, 2024. 2

    2024 arXiv

  37. [45]

    Instant volumetric head avatars

    Wojciech Zielonka, Timo Bolkart, and Justus Thies. Instant volumetric head avatars. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages 4574–4584, 2023. 2, 7 10

    2023

Pith tools

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