pith. sign in

arxiv: 2606.31115 · v1 · pith:4Z5E3IVRnew · submitted 2026-06-30 · 💻 cs.CV

JacobianAvatar: Temporally Consistent Semi-rigid Avatar Reconstruction from a Monocular Video

Changyeon Won , Min-Gyu Park , Seonghwan Park , Ju Hong Yoon , Hae-Gon Jeon This is my paper

Pith reviewed 2026-07-01 05:59 UTC · model grok-4.3

classification 💻 cs.CV
keywords avatar reconstructionmonocular videoneural Jacobian fieldssemi-rigid deformationtemporal consistencyPoisson equationhuman modelingself-supervised learning
0
0 comments X

The pith

Neural Jacobian fields solved via constrained Poisson equations reconstruct temporally consistent semi-rigid human avatars from monocular video.

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

The paper tries to establish a way to generate realistic human avatars that capture complex motions including clothing dynamics using only a single video feed. It represents deformations through neural networks that output Jacobian matrices, which are integrated by solving a Poisson equation in a self-supervised manner. Monocular capture introduces problems with hidden surfaces and motion jitter, so the work adds a constrained Poisson solver, signed-distance regularization on the Jacobians, and a residual flow loss guided by the deformations themselves. These additions aim to remove boundary problems, fill in areas like armpits and thighs, and keep the avatar coherent across frames, producing results that exceed prior methods on standard and real-world test videos.

Core claim

Neural Jacobian fields represent semi-rigid deformations by predicting Jacobian matrices whose integration is obtained by solving a Poisson equation; three added components—a constrained Poisson solver, signed distance-based Jacobian regularization, and deformation-guided residual flow loss—suppress boundary artifacts, recover frequently occluded regions such as armpits and thighs, and enforce temporal consistency during motion.

What carries the argument

Neural Jacobian fields (NJFs): self-supervised networks that predict pose-dependent Jacobian matrices, integrated through Poisson solving and regularized by the constrained solver, signed-distance term, and residual flow loss to handle monocular occlusions and motion.

If this is right

  • The reconstructed avatars exhibit temporal stability and geometric coherence across frames.
  • Occluded and invisible surfaces are recovered without additional views or sensors.
  • Boundary artifacts that appear in earlier monocular methods are reduced.
  • Performance exceeds state-of-the-art approaches on both benchmark datasets and unconstrained videos.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The same Jacobian representation and Poisson solving could be tested on non-human articulated objects if the deformation patterns are comparably semi-rigid.
  • The residual flow loss term might transfer to other neural-field deformation models that already use flow supervision.
  • If the method remains stable on longer sequences, it could support avatar creation pipelines that ingest casual phone footage without manual cleanup.

Load-bearing premise

The three introduced components are sufficient to suppress boundary artifacts and recover occluded regions such as armpits and thighs from monocular input alone.

What would settle it

Persistent boundary artifacts or visible failure to reconstruct occluded regions like armpits and thighs on benchmark or in-the-wild test sequences would show the components are not sufficient.

Figures

Figures reproduced from arXiv: 2606.31115 by Changyeon Won, Hae-Gon Jeon, Ju Hong Yoon, Min-Gyu Park, Seonghwan Park.

Figure 1
Figure 1. Figure 1: JacobianAvatar. A neural representation for digital human avatars that cap￾tures rigid articulated motions and non-rigid local deformations using hierarchical neu￾ral Jacobian fields, while encouraging temporal consistency with high-fidelity geometry. The top and bottom rows show the rendered color images and normal maps of an ani￾mated avatar. Abstract. Generating realistic human avatars in complex motion… view at source ↗
Figure 2
Figure 2. Figure 2: Overview of our pipeline. We first initialize our canonical avatar using a human template mesh and refine it through mesh optimization. Next, we model semi-rigid deformations using two Jacobian fields integrated with a screened Poisson solver, which separately capture deformations in a coarse-to-fine manner. For appearance, we model the mesh textures as normal-conditioned colors. Finally, we incorporate 3D… view at source ↗
Figure 3
Figure 3. Figure 3: Qualitative comparison of rendered normal maps. Compared methods [9, 43, 44] suffer from texture-copying artifacts on the geometry, whereas our method shows smooth surfaces in the textured regions. GT Image Ours Vid2Avatar LSAvatar Ours Vid2Avatar LSAvatar GT Image Ours Vid2Avatar LSAvatar Ours Vid2Avatar LSAvatar [PITH_FULL_IMAGE:figures/full_fig_p013_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Qualitative comparison of mesh reconstruction with SDF-based methods [9,43] on the DNA-Rendering dataset. Ground Truth Ours ExAvatar GoMAvatar LSAvatar Vid2Avatar FacAvatar [PITH_FULL_IMAGE:figures/full_fig_p013_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Qualitative comparison of rendering quality with state-of-the-art methods [9, 30, 43, 44, 50] on the MonoperfCap Dataset. extraction artifacts inherently caused by the Marching Cubes algorithm, which is essential for mesh reconstruction in previous SDF-based approaches. In [PITH_FULL_IMAGE:figures/full_fig_p013_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Ablation study on the Synwild dataset. We visualize the effects of each compo￾nent of our proposed method on the geometry quality by removing them independently [PITH_FULL_IMAGE:figures/full_fig_p014_6.png] view at source ↗
read the original abstract

Generating realistic human avatars in complex motions--such as clothing dynamics--requires modeling of global and local deformations which remains challenging in monocular settings. We address this problem by leveraging neural Jacobian fields (NJFs) for representing semi-rigid deformations. We train self-supervised neural networks for predicting Jacobian matrices that give the pose-dependent deformations, by solving a Poisson equation. However, monocular input presents several difficulties such as self-occluded regions and invisible surfaces. To address these issues, we introduce three key components: a constrained Poisson solver, signed distance-based Jacobian regularization, and a deformation-guided residual flow loss, which together suppress boundary artifacts, recover frequently occluded regions such as armpits and thighs, and enforce temporal consistency during motion. Experiments on benchmark and in-the-wild videos demonstrate that our method generates temporally stable and geometrically coherent avatars, outperforming state-of-the-art approaches.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

1 major / 0 minor

Summary. The paper proposes JacobianAvatar, a method for reconstructing semi-rigid avatars from monocular video using neural Jacobian fields (NJFs). It trains self-supervised networks to predict Jacobian matrices for pose-dependent deformations by solving a Poisson equation. Three key components are introduced: a constrained Poisson solver, signed distance-based Jacobian regularization, and a deformation-guided residual flow loss to handle self-occlusions, boundary artifacts, and temporal consistency. The method is evaluated on benchmark and in-the-wild videos, claiming to outperform state-of-the-art approaches in generating temporally stable and geometrically coherent avatars.

Significance. If the experimental validation holds, the work could advance monocular 3D human reconstruction by offering a self-supervised NJF-based approach that targets occlusions and temporal stability without multi-view input. The three proposed components represent targeted technical contributions to deformation modeling.

major comments (1)
  1. [Abstract] Abstract: the claim that the three components (constrained Poisson solver, signed distance-based Jacobian regularization, and deformation-guided residual flow loss) are sufficient to suppress boundary artifacts and recover occluded regions such as armpits and thighs from monocular input alone is presented without any quantitative results, error analysis, or derivation details, making it impossible to verify whether the math and data support the outperformance claim.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for the review and the opportunity to clarify our presentation. We address the major comment below.

read point-by-point responses

  1. Referee: [Abstract] Abstract: the claim that the three components (constrained Poisson solver, signed distance-based Jacobian regularization, and deformation-guided residual flow loss) are sufficient to suppress boundary artifacts and recover occluded regions such as armpits and thighs from monocular input alone is presented without any quantitative results, error analysis, or derivation details, making it impossible to verify whether the math and data support the outperformance claim.

    Authors: The abstract is a concise high-level summary, as is conventional. Quantitative results (including metrics on temporal consistency, geometric error, and ablation studies isolating each of the three components), error analyses, and mathematical derivations for the constrained Poisson solver and regularizers are provided in Sections 3 and 4 of the manuscript. These experiments on benchmark and in-the-wild data support the claims regarding artifact suppression and occluded-region recovery from monocular video. The outperformance statements are grounded in those results. revision: no

Circularity Check

0 steps flagged

No significant circularity detected

full rationale

The derivation proceeds from neural prediction of Jacobian matrices, followed by integration via a constrained Poisson solver plus explicit regularization terms (signed-distance Jacobian and deformation-guided flow losses) to handle monocular ambiguities. These steps are additive design choices rather than reductions of outputs back to inputs by definition; the self-supervised objective is constructed from geometric consistency constraints that remain independent of the final avatar geometry. No self-citation chain, fitted-parameter renaming, or ansatz smuggling is required for the core pipeline, and the empirical claims rest on external benchmark comparisons rather than internal redefinition.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Only the abstract is available, so no concrete free parameters, axioms, or invented entities can be extracted; the central claim rests on the unverified effectiveness of the three listed components for handling monocular occlusions.

pith-pipeline@v0.9.1-grok · 5695 in / 1129 out tokens · 27773 ms · 2026-07-01T05:59:19.145439+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

57 extracted references · 7 canonical work pages · 2 internal anchors

  1. [1]

    In: SIGGRAPH Asia 2023 Technical Communications

    Abdrashitov, R., Raichstat, K., Monsen, J., Hill, D.: Robust skin weights transfer via weight inpainting. In: SIGGRAPH Asia 2023 Technical Communications. SA ’23, Association for Computing Machinery, New York, NY, USA (2023).https: //doi.org/10.1145/3610543.3626180,https://doi.org/10.1145/3610543. 3626180

    work page doi:10.1145/3610543.3626180 2023

  2. [2]

    ACM TOG41(4) (2022)

    Aigerman, N., Gupta, K., Kim, V.G., Chaudhuri, S., Saito, J., Groueix, T.: Neural jacobian fields: learning intrinsic mappings of arbitrary meshes. ACM TOG41(4) (2022)

    2022

  3. [3]

    In: 2018 International Conference on 3D Vision (3DV)

    Alldieck, T., Magnor, M., Xu, W., Theobalt, C., Pons-Moll, G.: Detailed human avatars from monocular video. In: 2018 International Conference on 3D Vision (3DV). pp. 98–109. IEEE (2018) 16 C. Won et al

    2018

  4. [4]

    In: CVPR (2025)

    Chen, H., Peng, B., Tao, Y., Zhang, J.: Dˆ 3-human: Dynamic disentangled digital human from monocular video. In: CVPR (2025)

    2025

  5. [5]

    In: ICCV (2021)

    Chen, X., Zheng, Y., Black, M.J., Hilliges, O., Geiger, A.: SNARF: Differentiable forward skinning for animating non-rigid neural implicit shapes. In: ICCV (2021)

    2021

  6. [6]

    In: Proceedings of the IEEE/CVF International Conference on Computer Vision

    Cheng, W., Chen, R., Fan, S., Yin, W., Chen, K., Cai, Z., Wang, J., Gao, Y., Yu, Z., Lin, Z., et al.: Dna-rendering: A diverse neural actor repository for high- fidelity human-centric rendering. In: Proceedings of the IEEE/CVF International Conference on Computer Vision. pp. 19982–19993 (2023)

    2023

  7. [7]

    In: SIGGRAPH Asia 2022 Conference Papers

    Feng,Y.,Yang,J.,Pollefeys,M.,Black,M.J.,Bolkart,T.:Capturingandanimation of body and clothing from monocular video. In: SIGGRAPH Asia 2022 Conference Papers. pp. 1–9 (2022)

    2022

  8. [8]

    Ferguson, A., Osman, A.A.A., Bescos, B., Stoll, C., Twigg, C., Lassner, C., Otte, D., Vignola, E., Prada, F., Bogo, F., Santesteban, I., Romero, J., Zarate, J., Lee, J., Park, J., Yang, J., Doublestein, J., Venkateshan, K., Kitani, K., Kavan, L., Farra, M.D., Hu, M., Cioffi, M., Fabris, M., Ranieri, M., Modarres, M., Kadlecek, P., Khirodkar, R., Abdrashit...

    work page arXiv 2025

  9. [9]

    In: CVPR (2023)

    Guo, C., Jiang, T., Chen, X., Song, J., Hilliges, O.: Vid2avatar: 3d avatar re- construction from videos in the wild via self-supervised scene decomposition. In: CVPR (2023)

    2023

  10. [10]

    In: CVPR (June 2025)

    Guo, C., Li, J., Kant, Y., Sheikh, Y., Saito, S., Cao, C.: Vid2avatar-pro: Authentic avatar from videos in the wild via universal prior. In: CVPR (June 2025)

    2025

  11. [11]

    In: CVPR (2024)

    Ho, I., Song, J., Hilliges, O.: Sith: Single-view textured human reconstruction with image-conditioned diffusion. In: CVPR (2024)

    2024

  12. [12]

    In: CVPR (2024)

    Hu, L., Zhang, H., Zhang, Y., Zhou, B., Liu, B., Zhang, S., Nie, L.: GaussianAvatar: Towards realistic human avatar modeling from a single video via animatable 3D gaussians. In: CVPR (2024)

    2024

  13. [13]

    In: NeurIPS (2023)

    Huang, Y., Wang, J., Zeng, A., Cao, H., Qi, X., Shi, Y., Zha, Z.J., Zhang, L.: DreamWaltz: Make a scene with complex 3D animatable avatars. In: NeurIPS (2023)

    2023

  14. [14]

    In: CVPR (2021)

    Jafarian, Y., Park, H.S.: Learning high fidelity depths of dressed humans by watch- ing social media dance videos. In: CVPR (2021)

    2021

  15. [15]

    In: CVPR (2022)

    Jiang, B., Hong, Y., Bao, H., Zhang, J.: SelfRecon: Self reconstruction your digital avatar from monocular video. In: CVPR (2022)

    2022

  16. [16]

    In: CVPR (2023)

    Jiang, T., Chen, X., Song, J., Hilliges, O.: Instantavatar: Learning avatars from monocular video in 60 seconds. In: CVPR (2023)

    2023

  17. [17]

    In: ECCV (2022)

    Jiang, W., Yi, K.M., Samei, G., Tuzel, O., Ranjan, A.: NeuMan: Neural human radiance field from a single video. In: ECCV (2022)

    2022

  18. [18]

    ACM Transac- tions on Graphics (ToG)32(3), 1–13 (2013)

    Kazhdan, M., Hoppe, H.: Screened poisson surface reconstruction. ACM Transac- tions on Graphics (ToG)32(3), 1–13 (2013)

    2013

  19. [19]

    ACM TOG42(4) (2023)

    Kerbl, B., Kopanas, G., Leimkühler, T., Drettakis, G.: 3d gaussian splatting for real-time radiance field rendering. ACM TOG42(4) (2023)

    2023

  20. [20]

    In: ECCV (2024)

    Khirodkar, R., Bagautdinov, T., Martinez, J., Zhaoen, S., James, A., Selednik, P., Anderson, S., Saito, S.: Sapiens: Foundation for human vision models. In: ECCV (2024)

    2024

  21. [21]

    In: ECCV (2024)

    Kratimenos, A., Lei, J., Daniilidis, K.: Dynmf: Neural motion factorization for real-time dynamic view synthesis with 3d gaussian splatting. In: ECCV (2024)

    2024

  22. [22]

    ACM TOG39(6) (2020) JacobianAvatar 17

    Laine, S., Hellsten, J., Karras, T., Seol, Y., Lehtinen, J., Aila, T.: Modular primi- tives for high-performance differentiable rendering. ACM TOG39(6) (2020) JacobianAvatar 17

    2020

  23. [23]

    In: CVPR (2025)

    Lei, J., Weng, Y., Harley, A.W., Guibas, L., Daniilidis, K.: Mosca: Dynamic gaus- sian fusion from casual videos via 4d motion scaffolds. In: CVPR (2025)

    2025

  24. [24]

    In: CVPR (2025)

    Li, P., Zheng, W., Liu, Y., Yu, T., Li, Y., Qi, X., Chi, X., Xia, S., Cao, Y.P., Xue, W., Luo, W., Guo, Y.: Pshuman: Photorealistic single-image 3d human recon- struction using cross-scale multiview diffusion and explicit remeshing. In: CVPR (2025)

    2025

  25. [25]

    In: Proceedings of the IEEE/CVF international conference on computer vision

    Li, R., Yang, S., Ross, D.A., Kanazawa, A.: Ai choreographer: Music conditioned 3d dance generation with aist++. In: Proceedings of the IEEE/CVF international conference on computer vision. pp. 13401–13412 (2021)

    2021

  26. [26]

    In: CVPR (2024)

    Li, Z., Zheng, Z., Wang, L., Liu, Y.: Animatable gaussians: Learning pose- dependent gaussian maps for high-fidelity human avatar modeling. In: CVPR (2024)

    2024

  27. [27]

    In: CVPR (2024)

    Liu, X., Zhan, X., Tang, J., Shan, Y., Zeng, G., Lin, D., Liu, X., Liu, Z.: Human- Gaussian: Text-driven 3D human generation with gaussian splatting. In: CVPR (2024)

    2024

  28. [28]

    ACM TOG34(6) (Oct 2015)

    Loper, M., Mahmood, N., Romero, J., Pons-Moll, G., Black, M.J.: SMPL: A skinned multi-person linear model. ACM TOG34(6) (Oct 2015)

    2015

  29. [29]

    In: CVPR (2021)

    Ma, Q., Saito, S., Yang, J., Tang, S., Black, M.J.: SCALE: Modeling clothed hu- mans with a surface codec of articulated local elements. In: CVPR (2021)

    2021

  30. [30]

    In: ECCV (2024)

    Moon, G., Shiratori, T., Saito, S.: Expressive whole-body 3d gaussian avatar. In: ECCV (2024)

    2024

  31. [31]

    In: CVPR (June 2015)

    Newcombe,R.A.,Fox,D.,Seitz,S.M.:Dynamicfusion:Reconstructionandtracking of non-rigid scenes in real-time. In: CVPR (June 2015)

    2015

  32. [32]

    ACM Transactions on Graphics (TOG)40(6), 1–13 (2021)

    Nicolet, B., Jacobson, A., Jakob, W.: Large steps in inverse rendering of geometry. ACM Transactions on Graphics (TOG)40(6), 1–13 (2021)

    2021

  33. [33]

    In: CVPR (2019)

    Pavlakos, G., Choutas, V., Ghorbani, N., Bolkart, T., Osman, A.A.A., Tzionas, D., Black, M.J.: Expressive body capture: 3d hands, face, and body from a single image. In: CVPR (2019)

    2019

  34. [34]

    In: CVPR (2019)

    Pavlakos, G., Choutas, V., Ghorbani, N., Bolkart, T., Osman, A.A.A., Tzionas, D., Black, M.J.: Expressive body capture: 3D hands, face, and body from a single image. In: CVPR (2019)

    2019

  35. [35]

    In: CVPR (2024)

    Qian, Z., Wang, S., Mihajlovic, M., Geiger, A., Tang, S.: 3dgs-avatar: Animatable avatars via deformable 3d gaussian splatting. In: CVPR (2024)

    2024

  36. [36]

    Advances in Neural Infor- mation Processing Systems33, 22468–22478 (2020)

    Remelli, E., Lukoianov, A., Richter, S., Guillard, B., Bagautdinov, T., Baque, P., Fua, P.: Meshsdf: Differentiable iso-surface extraction. Advances in Neural Infor- mation Processing Systems33, 22468–22478 (2020)

    2020

  37. [37]

    In: CVPR (2021)

    Saito, S., Yang, J., Ma, Q., Black, M.J.: SCANimate: Weakly supervised learning of skinned clothed avatar networks. In: CVPR (2021)

    2021

  38. [38]

    In: ICCV (2025)

    Saleh, F., Aliakbarian, S., Hewitt, C., Petikam, L., Xiao, X., Criminisi, A., Cash- man, T.J., Baltrusaitis, T.: David: Data-efficient and accurate vision models from synthetic data. In: ICCV (2025)

    2025

  39. [39]

    In: CVPR (2024)

    Shao, Z., Wang, Z., Li, Z., Wang, D., Lin, X., Zhang, Y., Fan, M., Wang, Z.: Splattingavatar: Realistic real-time human avatars with mesh-embedded gaussian splatting. In: CVPR (2024)

    2024

  40. [40]

    In: ECCV (2024)

    Shin, J., Lee, J., Lee, S., Park, M.G., Kang, J.M., Yoon, J.H., Jeon, H.G.: Canoni- calFusion: Generating drivable 3D human avatars from multiple images. In: ECCV (2024)

    2024

  41. [41]

    In: ICCV (2025)

    Sim, G., Moon, G.: PERSONA: Personalized whole-body 3D avatar with pose- driven deformations from a single image. In: ICCV (2025)

    2025

  42. [42]

    arXiv preprint arXiv:2308.11951 (2023) 18 C

    Song, C., Wandt, B., Rhodin, H.: Pose modulated avatars from video. arXiv preprint arXiv:2308.11951 (2023) 18 C. Won et al

    work page arXiv 2023

  43. [43]

    In: The Thirteenth International Conference on Learning Rep- resentations (2025)

    Song, C., Wu, Z., Su, S.Y., Wandt, B., Sigal, L., Rhodin, H.: Locality sensitive avatars from video. In: The Thirteenth International Conference on Learning Rep- resentations (2025)

    2025

  44. [44]

    In: Computer Graphics Forum

    Song, C., Wu, Z., Wandt, B., Sigal, L., Rhodin, H.: Representing animatable avatar via factorized neural fields. In: Computer Graphics Forum. vol. 44, p. e70192. Wiley Online Library (2025)

    2025

  45. [45]

    In: ECCV

    Teed, Z., Deng, J.: Raft: Recurrent all-pairs field transforms for optical flow. In: ECCV. Springer (2020)

    2020

  46. [46]

    In: ICCV (2025)

    Wang, Q., Ye, V., Gao, H., Zeng, W., Austin, J., Li, Z., Kanazawa, A.: Shape of motion: 4d reconstruction from a single video. In: ICCV (2025)

    2025

  47. [47]

    ACM TOG2501.14726(2025)

    Wang, S., Simon, T., Santesteban, I., Bagautdinov, T., Li, J., Agrawal, V., Prada, F., Yu, S.I., Nalbone, P., Gramlich, M., Lubachersky, R., Wu, C., Romero, J., Saragih, J., Zollhoefer, M., Geiger, A., Tang, S., Saito, S.: Relightable full-body gaussian codec avatars. ACM TOG2501.14726(2025)

    work page arXiv 2025

  48. [48]

    arXiv preprint arXiv:2506.21526 (2025)

    Wang, Y., Deng, J.: Waft: Warping-alone field transforms for optical flow. arXiv preprint arXiv:2506.21526 (2025)

    work page arXiv 2025

  49. [49]

    In: ECCV

    Wang, Y., Lipson, L., Deng, J.: Sea-raft: Simple, efficient, accurate raft for optical flow. In: ECCV. Springer (2024)

    2024

  50. [50]

    In: CVPR (2024)

    Wen, J., Zhao, X., Ren, Z., Schwing, A.G., Wang, S.: Gomavatar: Efficient animat- able human modeling from monocular video using gaussians-on-mesh. In: CVPR (2024)

    2024

  51. [51]

    In: CVPR (2022)

    Weng, C.Y., Curless, B., Srinivasan, P.P., Barron, J.T., Kemelmacher-Shlizerman, I.: HumanNeRF: Free-viewpoint rendering of moving people from monocular video. In: CVPR (2022)

    2022

  52. [52]

    In: CVPR (2024)

    Wu,R.,Mildenhall,B.,Henzler,P.,Park,K.,Gao,R.,Watson,D.,Srinivasan,P.P., Verbin, D., Barron, J.T., Poole, B., Hoyński, A.: Reconfusion: 3d reconstruction with diffusion priors. In: CVPR (2024)

    2024

  53. [53]

    Structured 3D Latents for Scalable and Versatile 3D Generation

    Xiang, J., Lv, Z., Xu, S., Deng, Y., Wang, R., Zhang, B., Chen, D., Tong, X., Yang, J.: Structured 3d latents for scalable and versatile 3d generation. arXiv preprint arXiv:2412.01506 (2024)

    work page internal anchor Pith review Pith/arXiv arXiv 2024

  54. [54]

    ACM TOG37(2) (2018)

    Xu, W., Chatterjee, A., Zollhöfer, M., Rhodin, H., Mehta, D., Seidel, H.P., Theobalt, C.: Monoperfcap: Human performance capture from monocular video. ACM TOG37(2) (2018)

    2018

  55. [55]

    In: CVPR (2024)

    Zhang, J., Li, X., Zhang, Q., Cao, Y., Shan, Y., Liao, J.: Humanref: Single image to 3d human generation via reference-guided diffusion. In: CVPR (2024)

    2024

  56. [56]

    Hunyuan3D 2.0: Scaling Diffusion Models for High Resolution Textured 3D Assets Generation

    Zhao, Z., Lai, Z., Lin, Q., Zhao, Y., Liu, H., Yang, S., Feng, Y., Yang, M., Zhang, S., Yang, X., et al.: Hunyuan3d 2.0: Scaling diffusion models for high resolution textured 3d assets generation. arXiv preprint arXiv:2501.12202 (2025)

    work page internal anchor Pith review Pith/arXiv arXiv 2025

  57. [57]

    In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition

    Zhou, Y., Barnes, C., Lu, J., Yang, J., Li, H.: On the continuity of rotation rep- resentations in neural networks. In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition. pp. 5745–5753 (2019) JacobianAvatar 19 Supplementary Material A Demonstration of A vatar Animation We provide a supplementary video showcasing the animati...

    2019