Pith. sign in

REVIEW 4 major objections 6 minor 61 references

GaussianGAN: Real-Time Photorealistic controllable Human Avatars

T0 review · 4 major / 6 minor · reviewed 2026-08-05 · deepseek-v4-flash

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

Pith's one-line read GaussianGAN claims a two-stage Gaussian-splatting pipeline can synthesize photorealistic, controllable human avatars from a 3D skeleton alone, in real time, with less blur than prior NeRF and Gaussian baselines.

desk verdict Plausible real-time avatar pipeline with an unsupported SOTA claim; the ZJU 'novel pose' label is actually novel-view, and baselines aren't re-run. read the letter →

arxiv 2509.01681 v1 pith:3H5SYFXN submitted 2025-09-01 cs.CV

Mohamed Ilyes Lakhal , Richard Bowden This is my paper
classification cs.CV
keywords animatableavatars3DGaussiansplattingpointclouddensificationnovelviewsynthesisneuralrenderingsignlanguageproductionGANimage
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 tries to show that a real-time, photorealistic, controllable human avatar can be driven directly by a 3D skeleton, without relying on a skinned body model such as SMPL. The proposed pipeline first turns the skeleton into a dense point cloud by sampling cylinders around each limb, then deforms those points with learned residuals and renders both Gaussian-splat features and semantic segmentation maps into the target view. A UNet generator with a GAN and perceptual losses then turns those rendered modalities into a sharp image. On the ZJU-MoCap and THuman4 benchmarks the authors report pixel-fidelity scores above prior methods — 32.94 dB and 33.39 dB PSNR respectively — while rendering at 79 FPS. The same pipeline is applied to sign language production, where it synthesizes photorealistic novel poses of a signer.

What carries the argument

The load-bearing mechanism is the skeleton-to-Gaussian densification: Algorithm 1 creates a fixed-size point cloud by sampling 200 points uniformly on the surface of a cylinder built around each limb, with cylinder radius set to 0.1 times limb length, plus spheres and cuboids for face and torso. This fixed-topology point cloud, deformed by a learned residual MLP rather than by skinning, is what lets the pipeline accept arbitrary poses from a skeleton alone; differentiable Gaussian rasterization then renders both the learned appearance features and a separate novel-view semantic segmentation, and the UNet generator uses those modalities to synthesize the final photorealistic image.

What would settle it

Measure hand-region reconstruction error on held-out frames with interlaced fingers: if the cylinder-sampled Gaussians miss the interpenetrating geometry, the UNet must hallucinate the fingers, which should show up as a cluster of high LPIPS or low PSNR in the hand area even when global PSNR stays high.

Watch

Extended reading notes

Core claim

The central claim is that blur, the main weakness of NeRF- and Gaussian-based avatar renderers, can be removed by separating the problem into a geometric stage and an image-synthesis stage. From a skeleton p, Algorithm 1 builds an initial point cloud P by wrapping each limb in a cylinder, the face in a sphere, and the torso in a cuboid, and uniformly sampling points. These points initialize Gaussian primitives G, whose parameters are frozen; an MLP Ψ regresses a residual deformation ΔG that fits the Gaussians to the actual body shape for the given pose. The corrected Gaussians are rasterized into a feature map F, while a separate SMPL-based Gaussian segmentation module renders a semantic map

Load-bearing premise

A fixed-size point cloud sampled around skeleton limbs — deformed only by learned residual shifts, with no skinning — can supply enough geometry and texture to represent arbitrary poses, including tightly interlaced hands; anything the splats miss must be invented by the UNet.

Editorial extensions

If this is right

  • At 79 FPS, the full rendering pipeline is fast enough for interactive telepresence, live sign-language display, and real-time entertainment avatars, rather than offline synthesis only.
  • Because the initial geometry is built from a skeleton and a limb dictionary rather than from an SMPL mesh, the same recipe can in principle drive any articulated object with a known skeleton.
  • Injecting rendered semantic segmentation alongside splatted Gaussian features lets the generator focus on texture instead of geometry, which is why the ablation without S drops from 34.16 to 29.08 PSNR on subject 377.
  • The sign-language application suggests that neural rendering can replace 2D-skeleton post-processing for photorealistic signer synthesis, with the paper reporting 36.00 PSNR versus 30.26 for AnonySign and 30.09 for GauHuman on its captured data.

Reading between the lines

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

  • The fixed point-cloud size N is a tunable assumption: concentrating more sampled Gaussians on hands and face would likely close the gap on tightly interlaced fingers, at some cost to frame rate.
  • Because a UNet post-processes the splatted features, geometry that the sparse point cloud misses can be painted over in image space; this may make the system robust to skeleton-estimation error but could introduce temporal flicker in those hallucinated regions.
  • The novel-view segmentation module is not tied to SMPL; replacing the body template with an animal skeleton or a non-human character rig would let the identical pipeline synthesize photorealistic non-human avatars, a direction the paper hints at but does not test.
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 / 6 minor

Summary. The paper proposes GaussianGAN, a real-time controllable human avatar system that operates on a 3D skeleton input. It first densifies the skeleton into a point cloud by constructing cylinders around limbs, spheres for the face, and cuboids for the torso (Alg. 1). An MLP regresses residual Gaussian parameters to deform this fixed-topology point cloud, and differentiable Gaussian splatting renders feature maps. A separate novel-view segmentation module renders semantic maps using SMPL-based Gaussian splatting. A UNet generator, conditioned on the rendered features, segmentation, a GNN-encoded pose, and a 6D-rotation pose representation, produces the final photorealistic image. The authors claim real-time rendering at 79 FPS and state-of-the-art PSNR of 32.94 dB on ZJU-MoCap and 33.39 dB on THuman4, with an additional application to sign language production.

Significance. The core idea of combining skeleton-initialized 3D Gaussian splatting with a UNet generator and a novel-view segmentation modality is novel and clearly described. The two-stage training procedure (feature rendering then GAN-based image synthesis) is a plausible approach to high-frequency detail preservation. If the reported quantitative results were obtained under a unified evaluation protocol, the claimed gains (e.g., average 32.94 dB PSNR on ZJU-MoCap and 33.39 dB on THuman4) would be a meaningful advance, and the sign language production application is a valuable extension. However, the central 'state-of-the-art' claim is not currently supported by controlled comparisons, because baseline numbers appear to be taken from disparate papers without re-evaluation under a common protocol. The paper also ships no code, reports no error bars, and includes only a single-sequence ablation, all of which limit reproducibility and the strength of the empirical conclusions.

major comments (4)
  1. [Sec. IV-B, Tables I-II, Sec. IV-A] The claimed SOTA result rests on incomparable baseline numbers. The authors state in Sec. IV-B that 'there are no common datasets used by all methods in the literature. Therefore, we report two different sets of methods per dataset,' which indicates that the baseline metrics are taken from each method's original paper rather than re-evaluated under GaussianGAN's protocol. This is confirmed by the ZJU-MoCap setup in Sec. IV-A: 'we use one camera for training and the remaining 22 for testing.' This is a novel-view protocol for training poses, not the 'novel pose synthesis' claimed in the Table I caption. The model has seen each pose from the training camera, so the test is easier than a true novel-pose test. Similarly, for THuman4 the paper claims to follow PoseVocab's protocol but does not demonstrate that all listed baselines (PoseVocab, SLRF, TAVA, Ani-NeRF, ARAH, TexVocab, HuGS) used i
  2. [Table II caption] The caption for Table II reads: 'Our model achieves the best performance PSNR and the third best in SSIM and PSNR, respectively.' This is self-contradictory and ambiguous; likely one of the two occurrences of 'PSNR' should be 'LPIPS.' In addition, the LPIPS column lists integer values for baselines (17, 24, 32, 34, 37, 13.40, 19) and 17.97 for the proposed method. Without stating the scaling factor (e.g., percentage or ×1000), the numbers are not directly comparable, and the claim of 'third best in SSIM and LPIPS' is not verifiable from the table as printed.
  3. [Sec. IV-C, Table III] The ablation study is conducted on a single sequence (sequence 377 of ZJU-MoCap), so the component contributions (GCN, segmentation, feature loss) are not shown to generalize across subjects or datasets. Moreover, the text states: 'We can see that adding the graph convolutional network to encode the pose further improves the synthesis quality from 29.08 to 33.17 PSNR (Tab. III).' However, 29.08 is the PSNR for the 'w/o S' condition, not 'w/o GCN'; the actual GCN contribution is the difference between 33.17 (w/o GCN) and 34.16 (Full), i.e., 0.99 dB. The text misattributes the improvement and thus misrepresents the role of the GCN.
  4. [Sec. III-C] The method relies on a fixed-size point cloud without a skinning mechanism, where a learned residual field is expected to account for arbitrary articulated poses. This is a critical design assumption: the point cloud is initialized from cylinders/spheres/cuboids around a skeleton, and the MLP regresses residuals to deform it. The paper does not provide evidence that this fixed-topology representation can handle complex poses such as hand interlacement in the quantitative benchmarks. The sign language experiment uses a custom captured dataset with no release, and the qualitative results do not quantify difficulty. This assumption may limit generalizability and should be tested explicitly (e.g., on poses with severe self-contact) or discussed more candidly as a limitation.
minor comments (6)
  1. [Abstract] '32.94db' should be '32.94 dB.' Also, the average PSNR on ZJU-MoCap is not computed in Table I; the abstract reports 32.94 dB, but the table only gives per-subject numbers. Please state how the average is computed.
  2. [Sec. IV-C, Table IV] The text says 'We conducted our experiment using sequence 386 of the ZJU-MoCap dataset,' but Table IV reports mIoU for all six sequences (377, 386, 387, 392, 393, 394). Please correct the text or the table.
  3. [Alg. 2] The notation uses 'F' both for the rasterized features and for the UNet encoder-decoder generator, leading to confusing expressions such as 'F ← rasterise(...)' and 'Ipred ← F(F, S).' Use different symbols, e.g., Φ for the generator.
  4. [Sec. III-D] The LBS formula is written as x_o = LBS(x_c, {ΔB_b}, ω_{x_c}) = (Σ_b ω^b_{x_c} ΔB_b) x_c, which omits the bone transformation indices and the rotation/translation parts. This notation is imprecise and should be aligned with standard SMPL skinning notation.
  5. [Fig. 9] The caption lists 'EasyMocap [37]' but reference [37] is Neural Body (Peng et al.). If EasyMocap is intended, the reference is missing.
  6. [Sec. IV-D] The statement 'We also implement the novel-view segmentation module using the method described in [8]' is unclear: reference [8] is on hand-to-hand intersection resolution via occupancy networks, not a semantic segmentation rendering method. Please clarify how this method is used for hand segmentation.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: empirical training, held-out evaluation, no fit-as-prediction.

full rationale

GaussianGAN is an empirical learning system; its outputs are generated images and its inputs are skeletons, point clouds, segmentation maps, and pixel losses. There is no step where a 'prediction' is defined as the fitted value of an input. Alg. 1 constructs a fixed-size point cloud by sampling cylinders/spheres/cuboids around the skeleton; this is an initialization, not an output. The pose-guided deformation Ψ is trained with L1+SSIM against target images and therefore learns to make rendered features approximate the image—this is a standard conditional rendering/autoencoding objective, not a circular definition. The novel-view segmentation module renders SMPL vertex semantics via LBS and rasterization and is trained with cross-entropy against ground-truth segmentation; it conditions the generator and is not a renamed version of the predicted image. The generator is a UNet trained with reconstruction, perceptual, and adversarial losses on held-out cameras/poses. The paper contains self-citations ([8], [44], [45]) in the sign-language application as components/baselines, but the central avatar-rendering comparison (Tables I-II) does not reduce to those citations. Concerns about protocol comparability of baselines are correctness risks, not circularity. Therefore no circular step is evident.

Assumptions & free parameters 5 free parameters · 4 assumptions · 0 invented entities

The method rests on standard neural rendering components plus several hand-set hyperparameters. The most significant assumption is the fixed-topology point cloud from a skeleton, which lacks a skinning mechanism and is not ablated. The segmentation module also assumes reliable SMPL fits.

free parameters (5)
  • cylinder radius factor = 0.1
    Algorithm 1 sets r = 0.1 * limb length; this is set empirically without ablation.
  • points per limb = 200
    Algorithm 1 samples 200 points around each cylinder; chosen by hand.
  • points for face/torso = 500
    Sec. III-B: sphere/cuboid sampled with 500 points; chosen by hand.
  • LD-SSIM weight lambda = not specified (taken from [15])
    Loss in Eq. (6); the controlling factor is borrowed from prior work but remains a hyperparameter.
  • GAN generator learning rate = 2e-5
    Sec. IV: Adam learning rate for the generator; other components' learning rates are unspecified.
assumptions (4)
  • standard math Gaussian splatting formulation (equations 1-4) is taken as background from prior work [15].
    Sec. III-A uses the standard Gaussian splatting rendering equations without derivation.
  • domain assumption SMPL model and Linear Blend Skinning provide correct deformation for the novel-view segmentation module.
    Sec. III-D assumes SMPL fits are available and accurate, and that LBS deformation is sufficient for semantic segmentation.
  • domain assumption Mediapipe skeleton keypoints are accurate enough for point cloud initialization.
    The method starts from a 3D skeleton (Mediapipe), but the paper acknowledges depth ambiguity and does not solve it.
  • ad hoc to paper A fixed-size point cloud without skinning can represent arbitrary articulated poses through learned residual deformations.
    Sec. III-C freezes N and learns residual deformations via an MLP; no evidence that this topology covers detailed surfaces for complex poses.

how reviews work

0 comments
Cite this review

Pith. "Pith review of GaussianGAN: Real-Time Photorealistic controllable Human Avatars." pith.science (2026). https://pith.science/paper/3H5SYFXN

@misc{pith2026250901681,
  author       = {Pith},
  title        = {Pith review of: GaussianGAN: Real-Time Photorealistic controllable Human Avatars},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/3H5SYFXN}},
  note         = {Machine review of arXiv:2509.01681}
}
read the original abstract

Photorealistic and controllable human avatars have gained popularity in the research community thanks to rapid advances in neural rendering, providing fast and realistic synthesis tools. However, a limitation of current solutions is the presence of noticeable blurring. To solve this problem, we propose GaussianGAN, an animatable avatar approach developed for photorealistic rendering of people in real-time. We introduce a novel Gaussian splatting densification strategy to build Gaussian points from the surface of cylindrical structures around estimated skeletal limbs. Given the camera calibration, we render an accurate semantic segmentation with our novel view segmentation module. Finally, a UNet generator uses the rendered Gaussian splatting features and the segmentation maps to create photorealistic digital avatars. Our method runs in real-time with a rendering speed of 79 FPS. It outperforms previous methods regarding visual perception and quality, achieving a state-of-the-art results in terms of a pixel fidelity of 32.94db on the ZJU Mocap dataset and 33.39db on the Thuman4 dataset.

Figures

Figures reproduced from arXiv: 2509.01681 by the authors.

Figure 1
Figure 1. GaussianGAN takes a 3D human skeleton as input and then synthesises an image of a human from it. First, it reconstructs a dense point cloud by a novel densification strategy. Moreover, we propose a segmentation module that renders the semantics of the articulated object (e.g. human) in the target view. We use a UNet encoder-decoder model to synthesise the human in motion. Our GaussianGAN can be easily applied to var… view at source ↗
Figure 2
Figure 2. Point cloud initialisation: (a) input skeleton; (b) [PITH_FULL_IMAGE:figures/full_fig_p002_2.png] view at source ↗
Figure 3
Figure 3. Overview of our proposed GaussianGAN. Starting from a 3D skeleton p, representing a human in a certain posture. We create a point cloud based on the limbs following our proposed densification strategy as presented in Alg. 1. The initial Gaussian parameters G are formed from this point cloud. A pose-guided deformation module consisting of an MLP network regresses a residual ∆G of the parameters to better fit the huma… view at source ↗
Figures from the paper (7 more)
Figure 4
Figure 4. Figure 4: Novel-view segmentation module. We render novel￾view semantic segmentation using Gaussian splatting [15]. In particular, we use an SMPL-based approach to render the segmentation on the observation space given a target body pose θ. through a UNet [21] encoder-decoder mo…
Figure 5
Figure 5. Figure 5: High-frequency preservation. Our proposed model preserves the high frequency details of the body in the novel￾view and novel-pose compared to neural radiance fields approaches. We highlight this through canny edge detection. KEY – (left) GT; (middle) ours; (right) GauH…
Figure 7
Figure 7. Figure 7: Novel-view segmentation. Qualitative results of our proposed segmentation module (Sec. III-D) on novel-view and novel-pose sequences. Our module can accurately render novel view segmentation. KEY – (left) GT segmentation; (right) our segmentation results. TABLE V: SLP.…
Figure 6
Figure 6. Figure 6: Ablation. We highlight the importance of using different modalities to improve the synthesis quality. KEY – (right): generator w/o segmentation mask S; (middle): generator using S, but w/o GCN; (left): generator with both S and GCN [PITH_FULL_IMAGE:figures/full_fig_p0…
Figure 8
Figure 8. Figure 8: Qualitative comparison of our proposed GaussianGAN against a Gaussian splatting method (GauHuman [7]) and a NeRF approach (ARAH [52]). The results show novel-view test examples on the ZJU-MoCap dataset [36]. Ground-truth ours NV seg. (Sec. 3.4) GauHuman [7] EasyMocap […
Figure 9
Figure 9. Figure 9: Sign Language Production. Qualitative results comparing our method with two state-of-the-art solid approaches. We overlay the rendering of the hand-fitted from the triangulated skeleton in our Novel-view segmentation module (Sec. 3.4) and the fitted SMPL parameters wit…
Figure 10
Figure 10. Figure 10: GaussianGAN vs. neural rendering (only). We highlight some of the key features that our proposed model takes into account, namely blurriness and details preservation. KEY – (left) GauHuman [7]; (right) ours [PITH_FULL_IMAGE:figures/full_fig_p008_10.png]

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

61 extracted references · 57 canonical work pages

  1. [1]

    Belhumeur, D

    P. Belhumeur, D. Kriegman, and A. Yuille. The bas-relief ambiguity. In CVPR, pages 1060–1066, 1997

    work page 1997

  2. [2]

    A. Cao, C. Rockwell, and J. Johnson. FWD: Real-time Novel View Synthesis with Forward Warping and Depth. CVPR, 2022

    work page 2022

  3. [3]

    Y . Cao, Y . Ju, and D. Xu. 3DGS-DET: Empower 3D Gaussian Splatting with Boundary Guidance and Box-Focused Sampling for 3D Object Detection, 2024

    work page 2024

  4. [4]

    K. Gao, Y . Gao, H. He, D. Lu, L. Xu, and J. Li. Nerf: Neural radiance field in 3d vision, a comprehensive review, 2023

    work page 2023

  5. [5]

    I. J. Goodfellow, J. Pouget-Abadie, M. Mirza, B. Xu, D. Warde- Farley, S. Ozair, A. Courville, and Y . Bengio. Generative Adversarial Networks. In NeurIPS, 2014

    work page 2014

  6. [6]

    L. Hu, H. Zhang, Y . Zhang, B. Zhou, B. Liu, S. Zhang, and L. Nie. GaussianAvatar: Towards Realistic Human Avatar Modeling from a Single Video via Animatable 3D Gaussians. In CVPR, 2024

    work page 2024

  7. [7]

    Hu and Z

    S. Hu and Z. Liu. GauHuman: Articulated Gaussian Splatting from Monocular Human Videos. In CVPR, 2024

    work page 2024

  8. [8]

    Ivashechkin, O

    M. Ivashechkin, O. Mendez, and R. Bowden. Two Hands Are Better Than One: Resolving Hand to Hand Intersections via Occupancy Networks. In IEEE 18th International Conference on Automatic Face and Gesture Recognition (FG) , 2024

    work page 2024

Show all 61 references
  1. [9]

    D. L. James and C. D. Twigg. Skinning mesh animations. ACM TOG, 24(3):399–407, July 2005

    2005

  2. [10]

    R. Jena, G. S. Iyer, S. Choudhary, B. Smith, P. Chaudhari, and J. Gee. SplatArmor: Articulated Gaussian splatting for animatable humans from monocular RGB videos. arXiv preprint arXiv:2311.10812, 2023

    2023 arXiv

  3. [11]

    Jiang, M

    Y . Jiang, M. Habermann, V . Golyanik, and C. Theobalt. HiFECap: Monocular High-Fidelity and Expressive Capture of Human Perfor- mances. In British Machine Vision Conference (BMVC) , 2022

    2022

  4. [12]

    Johnson, A

    J. Johnson, A. Alahi, and L. Fei-Fei. Perceptual losses for real-time style transfer and super-resolution. In ECCV, pages 694–711, 2016

    2016

  5. [13]

    A. Kar, C. H ¨ane, and J. Malik. Learning a multi-view stereo machine. In Proceedings of the 31st International Conference on Neural Information Processing Systems , page 364–375, 2017

    2017

  6. [14]

    H. Kato, D. Beker, M. Morariu, T. Ando, T. Matsuoka, W. Kehl, and A. Gaidon. Differentiable rendering: A survey. ArXiv, abs/2006.12057, 2020

    2006 arXiv

  7. [15]

    Kerbl, G

    B. Kerbl, G. Kopanas, T. Leimk ¨uhler, and G. Drettakis. 3D Gaussian Splatting for Real-Time Radiance Field Rendering. ACM Transactions on Graphics, 42(4), July 2023

    2023

  8. [16]

    D. P. Kingma and J. Ba. Adam: A Method for Stochastic Optimization. In ICLR, 2015

    2015

  9. [17]

    Kocabas, J.-H

    M. Kocabas, J.-H. R. Chang, J. Gabriel, O. Tuzel, and A. Ranjan. HUGS: Human gaussian splatting. In CVPR, 2024

    2024

  10. [18]

    H. Le, T. Mensink, P. Das, and T. Gevers. Novel View Synthesis from Single Images via Point Cloud Transformation. In Proceedings of the British Machine Vision Conference (BMVC) , 2020

    2020

  11. [19]

    K. Lee, H. Chang, L. Jiang, H. Zhang, Z. Tu, and C. Liu. ViTGAN: Training GANs with Vision Transformers. In ICLR, 2022

    2022

  12. [20]

    R. Li, J. Tanke, M. V o, M. Zollhofer, J. Gall, A. Kanazawa, and C. Lassner. TA V A: Template-free animatable volumetric actors. In ECCV, 2022

    2022

  13. [21]

    X. Li, H. Chen, X. Qi, Q. Dou, C. Fu, and P. Heng. H-denseunet: Hybrid densely connected unet for liver and tumor segmentation from ct volumes. IEEE Transactions on Medical Imaging , 37(12):2663– 2674, 2018

    2018

  14. [22]

    Z. Li, Y . Sun, Z. Zheng, L. Wang, S. Zhang, and Y . Liu. Animatable and relightable gaussians for high-fidelity human avatar modeling. arXiv preprint arXiv:2311.16096v4 , 2024

    2024 arXiv

  15. [24]

    Z. Li, Z. Zheng, Y . Liu, B. Zhou, and Y . Liu. PoseV ocab: Learning Joint-structured Pose Embeddings for Human Avatar Modeling. In ACM SIGGRAPH Conference Proceedings , 2023

    2023

  16. [25]

    S. Liu, T. Li, W. Chen, and H. Li. Soft rasterizer: A differentiable renderer for image-based 3d reasoning. The IEEE International Conference on Computer Vision (ICCV) , Oct 2019

    2019

  17. [26]

    W. Liu, Z. Piao, M. Jie, L. M. Wenhan Luo, and S. Gao. Liquid Warping GAN: A Unified Framework for Human Motion Imitation, Appearance Transfer and Novel View Synthesis. In ICCV, 2019

    2019

  18. [27]

    Y . Liu, Z. Li, Y . Liu, and H. Wang. Texvocab: Texture vocabulary- conditioned human avatars. In Proceedings of the IEEE/CVF Con- ference on Computer Vision and Pattern Recognition (CVPR) , pages 1715–1725, June 2024

    2024

  19. [28]

    Loper, N

    M. Loper, N. Mahmood, J. Romero, G. Pons-Moll, and M. J. Black. SMPL: a skinned multi-person linear model. ACM TOG, 34(6), 2015

    2015

  20. [29]

    Lugaresi, J

    C. Lugaresi, J. Tang, H. Nash, C. McClanahan, E. Uboweja, M. Hays, F. Zhang, C. Chang, M. G. Yong, J. Lee, W. Chang, W. Hua, M. Georg, and M. Grundmann. Mediapipe: A framework for building perception pipelines. CoRR, abs/1906.08172, 2019

    1906 arXiv

  21. [30]

    Y . Men, Y . Mao, Y . Jiang, W.-Y . Ma, and Z. Lian. Controllable Person Image Synthesis with Attribute-Decomposed GAN. In CVPR, 2020

    2020

  22. [31]

    Mildenhall, P

    B. Mildenhall, P. P. Srinivasan, M. Tancik, J. T. Barron, R. Ramamoor- thi, and R. Ng. Nerf: Representing scenes as neural radiance fields for view synthesis. In ECCV, 2020

    2020

  23. [32]

    Moreau, J

    A. Moreau, J. Song, H. Dhamo, R. Shaw, Y . Zhou, and E. P ´erez- Pellitero. Human gaussian splatting: Real-time rendering of animatable avatars. In CVPR, 2024

    2024

  24. [33]

    Neverova, R

    N. Neverova, R. Alp G ¨uler, and I. Kokkinos. Dense pose transfer. In ECCV, 2018

    2018

  25. [34]

    H. Pang, H. Zhu, A. Kortylewski, C. Theobalt, and M. Habermann. ASH: Animatable Gaussian Splats for Efficient and Photoreal Human Rendering. In CVPR, pages 1165–1175, June 2024

    2024

  26. [35]

    S. Peng, J. Dong, Q. Wang, S. Zhang, Q. Shuai, X. Zhou, and H. Bao. Animatable Neural Radiance Fields for Modeling Dynamic Human Bodies. In ICCV, 2021

    2021

  27. [36]

    S. Peng, Y . Zhang, Y . Xu, Q. Wang, Q. Shuai, H. Bao, and X. Zhou. Neural body: Implicit neural representations with structured latent codes for novel view synthesis of dynamic humans. In CVPR, pages 9054–9063, June 2021

    2021

  28. [37]

    S. Peng, Y . Zhang, Y . Xu, Q. Wang, Q. Shuai, H. Bao, and X. Zhou. Neural Body: Implicit Neural Representations with Structured Latent Codes for Novel View Synthesis of Dynamic Humans. In CVPR, 2021

    2021

  29. [38]

    Z. Qian, S. Wang, M. Mihajlovic, A. Geiger, and S. Tang. 3DGS- Avatar: Animatable Avatars via Deformable 3D Gaussian Splatting. In CVPR, 2024

    2024

  30. [39]

    A. S. A. Rabby and C. Zhang. BeyondPixels: A Comprehensive Review of the Evolution of Neural Radiance Fields. arXiv e-prints , pages arXiv–2306, 2023

    2023

  31. [40]

    L. Ren, H. Pfister, and M. Zwicker. Object Space EW A Surface Splatting: A Hardware Accelerated Approach to High Quality Point Rendering. In Computer Graphics Forum / Eurographics , 2002

    2002

  32. [41]

    Rombach, A

    R. Rombach, A. Blattmann, D. Lorenz, P. Esser, and B. Ommer. High- resolution image synthesis with latent diffusion models. In CVPR, pages 10684–10695, June 2022

    2022

  33. [42]

    Rossi, M

    A. Rossi, M. Barbiero, P. Scremin, and R. Carli. Robust visibility surface determination in object space via pl ¨ucker coordinates. Journal of Imaging, 7(6), 2021

    2021

  34. [43]

    Saito, G

    S. Saito, G. Schwartz, T. Simon, J. Li, and G. Nam. Relightable Gaussian Codec Avatars. In CVPR, 2024

    2024

  35. [44]

    Saunders, N

    B. Saunders, N. C. Camgoz, and R. Bowden. Anonysign: Novel Hu- man Appearance Synthesis for Sign Language Video Anonymisation. In 2021 16th IEEE International Conference on Automatic Face and Gesture Recognition (FG 2021) , 2021

    2021

  36. [45]

    Saunders, N

    B. Saunders, N. C. Camgoz, and R. Bowden. Signing at Scale: Learning to Co-Articulate Signs for Large-Scale Photo-Realistic Sign Language Production. In CVPR, 2022

    2022

  37. [46]

    J. L. Sch ¨onberger and J.-M. Frahm. Structure-from-Motion Revisited. In CVPR, 2016

    2016

  38. [47]

    Z. Shao, Z. Wang, Z. Li, D. Wang, X. Lin, Y . Zhang, M. Fan, and Z. Wang. SplattingAvatar: Realistic Real-Time Human Avatars with Mesh-Embedded Gaussian Splatting. In CVPR, 2024

    2024

  39. [48]

    S. Song, W. Zhang, J. Liu, and T. Mei. Unsupervised Person Image Generation with Semantic Parsing Transformation. In CVPR, 2019

    2019

  40. [49]

    S.-Y . Su, T. Bagautdinov, and H. Rhodin. DANBO: Disentangled Articulated Neural Body Representations via Graph Neural Networks. In ECCV, 2022

    2022

  41. [50]

    Svitov, P

    D. Svitov, P. Morerio, L. Agapito, and A. D. Bue. HAHA: Highly Articulated Gaussian Human Avatars with Textured Mesh Prior, 2024

    2024

  42. [51]

    M. A. Uy, R. Martin-Brualla, L. Guibas, and K. Li. SCADE: NeRFs from Space Carving with Ambiguity-Aware Depth Estimates. In CVPR, 2023

    2023

  43. [52]

    S. Wang, K. Schwarz, A. Geiger, and S. Tang. ARAH: Animatable V olume Rendering of Articulated Human SDFs. In ECCV, 2022

    2022

  44. [53]

    Z. Wang, A. C. Bovik, H. R. Sheikh, and E. P. Simoncelli. Image Quality Assessment: From Error Visibility to Structural Similarity. IEEE TIP, 13(4):600–612, 2004

    2004

  45. [54]

    J. Wen, X. Zhao, Z. Ren, A. Schwing, and S. Wang. GoMAvatar: Efficient Animatable Human Modeling from Monocular Video Using Gaussians-on-Mesh. In CVPR, 2024

    2024

  46. [55]

    C.-Y . Weng, B. Curless, P. P. Srinivasan, J. T. Barron, and I. Kemelmacher-Shlizerman. HumanNeRF: Free-viewpoint rendering of moving people from monocular video. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 16210–16220, June 2022

    2022

  47. [56]

    Z. Yu, W. Cheng, X. Liu, W. Wu, and K.-Y . Lin. MonoHuman: Animatable Human Neural Field From Monocular Video. In Proceed- ings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 16943–16953, June 2023

    2023

  48. [57]

    Yunus, J

    R. Yunus, J. E. Lenssen, M. Niemeyer, Y . Liao, C. Rupprecht, C. Theobalt, G. Pons-Moll, J.-B. Huang, V . Golyanik, and E. Ilg. Recent Trends in 3D Reconstruction of General Non-Rigid Scenes. Computer Graphics Forum, 43(2), 2024

    2024

  49. [58]

    Zhang, P

    R. Zhang, P. Isola, A. A. Efros, E. Shechtman, and O. Wang. The Unreasonable Effectiveness of Deep Features as a Perceptual Metric. In CVPR, 2018

    2018

  50. [59]

    Zheng, H

    Z. Zheng, H. Huang, T. Yu, H. Zhang, Y . Guo, and Y . Liu. Structured Local Radiance Fields for Human Avatar Modeling. In CVPR, June 2022

    2022

  51. [60]

    X. Zhou, Z. Lin, X. Shan, Y . Wang, D. Sun, and M.-H. Yang. DrivingGaussian: Composite Gaussian Splatting for Surrounding Dy- namic Autonomous Driving Scenes. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) , pages 21634–21643, June 2024

    2024

  52. [61]

    Y . Zhou, C. Barnes, J. Lu, J. Yang, and H. Li. On the Continuity of Rotation Representations in Neural Networks. In CVPR, June 2019

    2019

  53. [62]

    Zuffi, A

    S. Zuffi, A. Kanazawa, and M. J. Black. Lions and tigers and bears: Capturing non-rigid, 3D, articulated shape from images. In CVPR, 2018

    2018

Pith tools

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