Pith. sign in

REVIEW 3 major objections 5 minor 71 references

Reconstructing Close Human Interaction with Appearance and Proxemics Reasoning

T0 review · 3 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 Reconstructing close human interactions from monocular video by jointly optimizing motion and appearance.

desk verdict Solid, honest empirical package for close two-person reconstruction, but the paper's central appearance-reasoning claim is not yet proven; the rest of the pipeline likely carries the results. read the letter →

arxiv 2507.02565 v1 pith:6A5J4WFU submitted 2025-07-03 cs.CV

Buzhen Huang , Chen Li , Chongyang Xu , Dongyue Lu , Jinnan Chen , Yangang Wang , Gim Hee Lee This is my paper
classification cs.CV
keywords 3Dhumanposeestimationcloseinteractiondual-branchoptimizationdiffusionproxemicspriorGaussiansplattingappearancereasoningin-the-wildreconstructionpseudoground-truthdataset
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

Close two-person interactions are hard to reconstruct from a single video because occluded bodies create depth ambiguities that even modern segmentation and keypoint models fail to resolve. This paper proposes a dual-branch optimization framework that reconstructs the two people's body motions and, at the same time, a coarse per-video 3D Gaussian appearance of each person. The core claim is that rendering the reconstructed Gaussians into the image and comparing them with the raw RGB frames gives a trustworthy depth-ordinal and alignment signal, so the optimization can decide who is in front and how the bodies touch. The motion side is constrained by a diffusion model trained on large indoor interaction datasets to provide proxemics and pose priors that transfer to outdoor scenes. On Hi4D and 3DPW the method reports the best MPJPE among the compared approaches (59.1 mm and 64.5 mm), and the authors release a 100-video in-the-wild interaction dataset with pseudo ground truth built from their framework.

What carries the argument

The load-bearing object is the dual-branch optimization. The motion branch is a pre-trained diffusion model with cross-attention between the two people; during optimization its network parameters are finetuned, so the prior itself regularizes the predicted motion. The appearance branch is an optimizable latent tensor decoded by a U-Net into a 14-channel Gaussian UV map, mapped onto the SMPL body surface and rendered with 3D Gaussian Splatting; an identity channel restricts keypoint supervision to each individual's rendered region. The two branches are optimized jointly against an appearance loss (L1, SSIM, LPIPS), a keypoint reprojection loss, a mesh-penetration distance field, and a smoothness term. This rendering-and-compare loop is the machinery that turns raw RGB into depth-ordinal constraints on pose.

What would settle it

Take a close-interaction video in which the two people wear identical clothing, and render the two reconstructed bodies with swapped appearance textures; if the method still recovers correct depth ordering and pose accuracy, the appearance branch is not the source of the disambiguation, and the central claim would be unsupported.

Watch

Extended reading notes

Core claim

The paper claims that jointly optimizing a diffusion-based motion prior and a per-video 3D Gaussian appearance model, under appearance, proxemics, and physics constraints, recovers accurate 3D close interactions from monocular in-the-wild videos, and that this outperforms existing single- and multi-person reconstruction approaches. The central mechanism is that the appearance branch converts raw RGB into pose constraints: splatting both people's Gaussians to the image plane lets the optimization compare rendered and original images, which disambiguates depth ordering and improves model-image alignment in occluded frames. The diffusion prior contributes by keeping the non-convex optimization away from implausible local minima, while a differentiable penetration loss keeps the bodies physically plausible.

Load-bearing premise

The load-bearing premise is that the appearance branch—two optimizable tensors decoded by a per-video U-Net into Gaussian UV maps—can be optimized per video, without pretraining, so that its RGB rendering loss genuinely constrains body pose rather than merely overfitting the image regardless of pose.

Editorial extensions

If this is right

  • On the two benchmarks tested, the method reports the lowest MPJPE among the compared methods (59.1 mm on Hi4D and 64.5 mm on 3DPW), indicating that appearance-based constraints improve pose accuracy in addition to depth ordering.
  • Because the framework needs only a whole-scene foreground mask rather than per-person segmentation, it sidesteps the failure of current semantic-parsing models in close contact.
  • The diffusion proxemics prior, trained on indoor interaction motion data, transfers to outdoor in-the-wild videos, making optimization-based reconstruction usable outside studios.
  • The pseudo ground-truth annotations produced by the framework improve a regression-based interaction method when used as training data, suggesting a data-generation loop.
  • When the appearance branch is removed, depth ordinal relationships become incorrect even though body poses stay accurate, which isolates the appearance cue's role.

Reading between the lines

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

  • The same dual-branch scheme could extend to crowds of more than two people, since the appearance branch only needs an aggregate foreground mask; the main obstacle is training the proxemics prior on multi-person interaction data.
  • At roughly 3–5 minutes per 128-frame video, the optimization is fast enough that faster Gaussian splatting renderers could turn it into an interactive annotation tool.
  • Adversarial appearance settings, such as mismatched or camouflage textures between the two people, should degrade depth-ordering accuracy if the appearance branch is doing real work, which gives a direct test of the mechanism.
  • The released pseudo ground-truth dataset could be used to train feed-forward regression models, potentially closing the loop between slow optimization quality and fast inference.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 5 minor

Summary. The paper proposes a dual-branch optimization framework for reconstructing two-person close interactions from monocular in-the-wild video. The motion branch is a diffusion model trained on interaction datasets to encode pose and proxemics priors; the appearance branch uses per-video optimizable tensors decoded by a U-Net into Gaussian UV maps, which are splatted and compared with the input RGB image through L1, SSIM, and LPIPS losses. The full objective combines appearance, keypoint reprojection, penetration, smoothness, and regularization terms, and both branches are optimized jointly. The authors also introduce WildCHI, a dataset of 100 TikTok videos with pseudo ground-truth annotations generated by the proposed method, and show that training CloseInt on this dataset improves its performance. Experiments on Hi4D and 3DPW report state-of-the-art MPJPE values (59.1 and 64.5, respectively).

Significance. If the main claim holds, the paper makes a useful contribution: it provides evidence that differentiable rendering of a coarse appearance model can help disambiguate depth ordering in close human interactions, a setting where keypoints and masks are unreliable. The framework is clearly described and compared against relevant recent baselines (BUDDI, CloseInt, BEV, GroupRec), and the ablation study covers the main components. The promise of code and data is also a strength. The main caveat is that the central appearance-pose coupling is validated only indirectly, and the WildCHI pseudo-labeling is self-referential without independent verification. As a result, the significance depends on additional controlled experiments that are within reach.

major comments (3)
  1. [Section 3.3, Eq. (6), Table 3] The central claim that the appearance branch constrains body pose through rendering is supported only by a 1.6 mm MPJPE gap (Ours w/o Appearance 60.68 vs Ours 59.06), which is smaller than other ablation differences in the same table (e.g., Ours w/o Physics reaches 57.01) and is reported from single runs without error bars. Since the appearance branch is re-optimized per video and can adjust Gaussian opacity, color, and offsets independently of the SMPL pose, it could in principle fit the RGB image without moving the body. Please add controlled experiments that separate pose sensitivity from photometric overfitting: (i) optimize the appearance branch with fixed ground-truth poses and with fixed perturbed poses and report the final Lapp values; (ii) run the full optimization with the appearance branch frozen and compare against the full model; (iii) report means and standard deviations over at least three random seeds for Table 3. Without such evidence, the 'appearance resolves depth ambiguity' contribution is not established.
  2. [Section 4 and Table 2 (WildCHI)] The pseudo ground-truth annotations in WildCHI are generated by the proposed method itself, and the only validation that the dataset improves regression-based models is training CloseInt on those pseudo-GT labels. This is self-referential unless the pseudo-GT accuracy is independently quantified. Please provide (i) quantitative agreement of the pseudo-GT with manually corrected annotations on a random subset, (ii) an evaluation that transfers a model trained on WildCHI to Hi4D/3DPW, or (iii) a comparison with pseudo-GT generated by an independent method such as BUDDI. This is needed to support the dataset contribution, although it is not necessary for the validity of the Hi4D/3DPW comparisons.
  3. [Section 3.3, Eq. (10), and Section 5.4] The overall objective is a weighted sum of at least six loss terms, but the loss weights are not reported. The paper also does not specify the diffusion training schedule, the mask strategy probabilities, the latent tensor dimensionality, or the U-Net architecture. Because the method is optimization-based and the reported numbers depend directly on these choices, the experiments are not reproducible from the text alone. Please include the full configuration (weights for Lapp, Lreproj, Lpen, Lsmooth, Lreg, optimizer schedules, and all hyperparameters) or point to released code with the exact settings used to produce Tables 2 and 3.
minor comments (5)
  1. [Table 2] The row label 'CLLIF' should be 'CLIFF' to match Reference [31].
  2. [Section 5.1] The phrase 'proxemixs prior' contains a typo and should read 'proxemics prior'.
  3. [Section 5.2] The 'interaction error' metric is mentioned but never defined; unless it is taken from CloseInt [16], please provide the exact formula or a precise reference.
  4. [Section 3.3, Eq. (7)] The word 'differentialable' is a typo, and the notation na and nb for face normals should be defined explicitly.
  5. [Section 3.3, appearance branch] The identity channel d in the Gaussian UV map is used to restrict keypoint supervision to the rendered region of each individual, but the mechanism is only described vaguely; please clarify how the two-person Gaussian splatting attributes pixels to a specific person's rendered region.

Circularity Check

1 steps flagged · score 2.0 of 10

Main held-out benchmark comparisons are not circular; the only notable self-reference is the WildCHI pseudo-GT, which is generated by the proposed method and then used to validate the dataset's value.

  1. other [Section 4 (WildCHI Dataset), first paragraph; Table 2]
    "we collect 100 videos with diverse environments and subjects from TikTok [54], and build pseudo ground-truth with the proposed method. We also manually filter out incorrect estimations. ... We also train CloseInt [16], a regression-based close human interaction method, on the proposed dataset. The experimental results in Tab. 2 show that WildCHI can improve the performance of CloseInt [16] in both indoor and outdoor scenarios."

    The pseudo ground-truth used to demonstrate the value of WildCHI is itself generated by the proposed method. Training CloseInt on labels produced by the same system and reporting improved downstream metrics shows that the student can learn from the teacher's outputs, but it does not independently validate the pseudo-GT accuracy: the proposed method is both the annotator and the source of the labels being evaluated. This is a self-referential validation of the dataset contribution. It is not a tautology (the student could fail to improve), and the paper's main comparisons on Hi4D and 3DPW use held-out ground truth, so this circularity is secondary rather than load-bearing for the central reconstruction claims.

full rationale

The central scientific claim — that jointly optimizing motion, appearance, proxemics, and physics reconstructs close interactions more accurately — is evaluated against held-out ground truth on Hi4D and 3DPW, with external baselines (Human4D, CLIFF, BEV, GroupRec, BUDDI, CloseInt) and component ablations. The appearance branch is a per-video analysis-by-synthesis fit; while it could in principle overfit the RGB input without constraining pose, that is an empirical vulnerability and not a definitional circularity because the pose is still scored against external GT. The diffusion proxemics prior is trained on separate interaction datasets (Inter-X, InterHuman) and fine-tuned at test time, which is a standard fitting strategy rather than a self-citation chain. The only material self-reference is the WildCHI pseudo-GT: it is generated by the proposed method and then used to show the dataset improves CloseInt. This weakens the dataset-validation evidence but does not undermine the main held-out benchmark results. Overall, the paper's core derivation is self-contained against external benchmarks, with a moderate but secondary self-referential dataset validation.

Assumptions & free parameters 4 free parameters · 5 assumptions · 2 invented entities

The paper is an empirical optimization method, so the ledger is dominated by hand-chosen hyperparameters and domain assumptions rather than derived constants. The most consequential entries are the unspecified loss weights and appearance-branch capacity, which together determine whether the RGB constraint actually supervises pose.

free parameters (4)
  • Loss weights in Eq. 10 = not reported
    The overall objective L = Lapp + Lreproj + Lpen + Lsmooth + Lreg presumably weights each term; no coefficients are listed, and the ablation table suggests they were tuned on Hi4D.
  • Optimization schedule = LR 2e-5 (motion), 3e-3 (appearance); steps not reported
    Adam learning rates are given in Sec. 3.3, but the number of iterations, warmup, and early stopping are not; these affect the balance between prior preservation and data fitting.
  • Appearance latent tensor dimensionality and U-Net architecture = not reported
    The latent code size, UV map resolution, and U-Net depth and channels are unspecified, yet they determine the capacity of the appearance branch and therefore the strength of the RGB constraint.
  • Mask strategy probabilities = not reported
    The diffusion prior uses two random masking schemes, temporal and counterpart masking, in Sec. 3.2; the masking rates are not given, affecting the learned prior's robustness behavior.
assumptions (5)
  • standard math SMPL with 6D rotation is a sufficient representation of two-person interaction.
    Adopted in Sec. 3.1 from cited works [36, 71].
  • domain assumption A perspective camera with a 55 degree diagonal FOV captures the in-the-wild image geometry.
    Follows CLIFF [31] in Sec. 3.2; if the true FOV differs, absolute translations will be biased.
  • domain assumption Diffusion prior trained on motion datasets without RGB (Inter-X, InterHuman) transfers to real videos when image features are zeroed.
    Sec. 3.2 states image features are set to zero for non-RGB data; the prior must generalize to the image-conditioned test distribution.
  • domain assumption The input video contains frames with little or no contact so that appearance can be initialized.
    Stated in the Limitation section as a requirement for the method to work.
  • ad hoc to paper Pseudo ground-truth annotations in WildCHI are accurate after manual filtering.
    Sec. 4 generates WildCHI with the proposed method, then manually filters out incorrect estimations; no independent verification protocol is given.
invented entities (2)
  • Per-person optimizable appearance latent tensor
    purpose: Drives the U-Net to regress a Gaussian UV map encoding color, opacity, rotation, scale, and identity for each person.
    Internal to the method; no falsifiable prediction outside the paper.
  • Identity channel d in the Gaussian UV map
    purpose: Labels each rendered Gaussian with person identity so keypoints can be assigned to the correct individual during optimization.
    A bookkeeping channel defined in Sec. 3.3; no external handle.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Reconstructing Close Human Interaction with Appearance and Proxemics Reasoning." pith.science (2026). https://pith.science/paper/6A5J4WFU

@misc{pith2026250702565,
  author       = {Pith},
  title        = {Pith review of: Reconstructing Close Human Interaction with Appearance and Proxemics Reasoning},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/6A5J4WFU}},
  note         = {Machine review of arXiv:2507.02565}
}
read the original abstract

Due to visual ambiguities and inter-person occlusions, existing human pose estimation methods cannot recover plausible close interactions from in-the-wild videos. Even state-of-the-art large foundation models~(\eg, SAM) cannot accurately distinguish human semantics in such challenging scenarios. In this work, we find that human appearance can provide a straightforward cue to address these obstacles. Based on this observation, we propose a dual-branch optimization framework to reconstruct accurate interactive motions with plausible body contacts constrained by human appearances, social proxemics, and physical laws. Specifically, we first train a diffusion model to learn the human proxemic behavior and pose prior knowledge. The trained network and two optimizable tensors are then incorporated into a dual-branch optimization framework to reconstruct human motions and appearances. Several constraints based on 3D Gaussians, 2D keypoints, and mesh penetrations are also designed to assist the optimization. With the proxemics prior and diverse constraints, our method is capable of estimating accurate interactions from in-the-wild videos captured in complex environments. We further build a dataset with pseudo ground-truth interaction annotations, which may promote future research on pose estimation and human behavior understanding. Experimental results on several benchmarks demonstrate that our method outperforms existing approaches. The code and data are available at https://www.buzhenhuang.com/works/CloseApp.html.

Figures

Figures reproduced from arXiv: 2507.02565 by the authors.

Figure 1
Figure 1. Due to the visual ambiguity, even state-of-the-art vision foundation models ( [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. With predicted UV Gaussian maps, we can map the Gaussians to 3D space with a UV coordinate map and splat them to the image [PITH_FULL_IMAGE:figures/full_fig_p002_2.png] view at source ↗
Figure 3
Figure 3. Overview of our framework. We propose a dual-branch optimization framework to reconstruct close human interactions from a monocular in-the-wild video. By optimizing the proxemics prior, U-Net backbone, and two optimizable tensors, the framework simul￾taneously predicts interactive motions and coarse appearances. With the constraints from 2D observations, physics, and prior knowledge, the framework can finally output… view at source ↗
Figures from the paper (2 more)
Figure 4
Figure 4. Figure 4: Qualitative comparison with BUDDI [40] and CloseInt [16]. Our method is more robust to visual ambiguity. Method Hi4D 3DPW MPJPE PA-MPJPE MPVPE Inter MPJPE PA-MPJPE MPVPE Inter Human4D [9] 72.1 52.4 88.6 – 72.9 49.1 81.8 – CLLIF [31] 91.3 53.6 109.6 141.5 – – – – BEV [5…
Figure 5
Figure 5. Figure 5: Ablation study. The initial prediction is severely affected by visual ambiguity and cannot reconstruct accurate interaction. With [PITH_FULL_IMAGE:figures/full_fig_p008_5.png]

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

71 extracted references · 63 canonical work pages

  1. [1]

    Federica Bogo, Angjoo Kanazawa, Christoph Lassner, Peter Gehler, Javier Romero, and Michael J. Black. Keep it SMPL: Automatic estimation of 3D human pose and shape from a single image. In ECCV, 2016. 3

    work page 2016

  2. [2]

    Keep it smpl: Automatic estimation of 3d human pose and shape from a single image

    Federica Bogo, Angjoo Kanazawa, Christoph Lassner, Peter Gehler, Javier Romero, and Michael J Black. Keep it smpl: Automatic estimation of 3d human pose and shape from a single image. In ECCV, pages 561–578, 2016. 3, 5

    work page 2016

  3. [3]

    Multi-person 3d pose and shape es- timation via inverse kinematics and refinement

    Junuk Cha, Muhammad Saqlain, GeonU Kim, Mingyu Shin, and Seungryul Baek. Multi-person 3d pose and shape es- timation via inverse kinematics and refinement. In ECCV, pages 660–677, 2022. 3

    work page 2022

  4. [4]

    Learning to estimate robust 3d human mesh from in-the-wild crowded scenes

    Hongsuk Choi, Gyeongsik Moon, JoonKyu Park, and Ky- oung Mu Lee. Learning to estimate robust 3d human mesh from in-the-wild crowded scenes. In CVPR, pages 1475– 1484, 2022. 2

    work page 2022

  5. [5]

    Monocular 3d reconstruction of interacting hands via collision-aware factorized refinements

    Rong et al. Monocular 3d reconstruction of interacting hands via collision-aware factorized refinements. In 3DV, 2021. 6

    work page 2021

  6. [6]

    Capturing closely in- teracted two-person motions with reaction priors

    Qi Fang, Yinghui Fan, Yanjun Li, Junting Dong, Dingwei Wu, Weidong Zhang, and Kang Chen. Capturing closely in- teracted two-person motions with reaction priors. In CVPR,

    work page

  7. [7]

    Three- dimensional reconstruction of human interactions

    Mihai Fieraru, Mihai Zanfir, Elisabeta Oneata, Alin-Ionut Popa, Vlad Olaru, and Cristian Sminchisescu. Three- dimensional reconstruction of human interactions. In CVPR, pages 7214–7223, 2020. 3, 6

    work page 2020

  8. [8]

    Reconstruct- ing three-dimensional models of interacting humans

    Mihai Fieraru, Mihai Zanfir, Elisabeta Oneata, Alin-Ionut Popa, Vlad Olaru, and Cristian Sminchisescu. Reconstruct- ing three-dimensional models of interacting humans. arXiv preprint arXiv:2308.01854, 2023. 3

    arXiv 2023

Show all 71 references
  1. [9]

    Humans in 4d: Re- constructing and tracking humans with transformers

    Shubham Goel, Georgios Pavlakos, Jathushan Rajasegaran, Angjoo Kanazawa, and Jitendra Malik. Humans in 4d: Re- constructing and tracking humans with transformers. In ICCV, pages 14783–14794, 2023. 1, 2, 3, 4, 6, 7

    2023

  2. [10]

    Multi-person extreme motion pre- diction

    Wen Guo, Xiaoyu Bie, Xavier Alameda-Pineda, and Francesc Moreno-Noguer. Multi-person extreme motion pre- diction. In CVPR, pages 13053–13064, 2022. 6

    2022

  3. [11]

    Panoptic studio: A massively multiview system for social interaction capture

    TomasSimon HanbyulJoo, HaoLiu XulongLi, LinGui Lei- Tan, and TimothyGodisart SeanBanerjee. Panoptic studio: A massively multiview system for social interaction capture. TPAMI, 41(1), 2019. 3

    2019

  4. [12]

    Gaussianavatar: Towards realistic human avatar modeling from a single video via animatable 3d gaussians

    Liangxiao Hu, Hongwen Zhang, Yuxiang Zhang, Boyao Zhou, Boning Liu, Shengping Zhang, and Liqiang Nie. Gaussianavatar: Towards realistic human avatar modeling from a single video via animatable 3d gaussians. In CVPR,

  5. [13]

    Gauhuman: Articulated gaus- sian splatting from monocular human videos

    Shoukang Hu and Ziwei Liu. Gauhuman: Articulated gaus- sian splatting from monocular human videos. In CVPR,

  6. [14]

    Pose2uv: Single-shot multiperson mesh recovery with deep uv prior

    Buzhen Huang, Tianshu Zhang, and Yangang Wang. Pose2uv: Single-shot multiperson mesh recovery with deep uv prior. TIP, 31:4679–4692, 2022. 3

    2022

  7. [15]

    Reconstructing groups of people with hypergraph relational reasoning

    Buzhen Huang, Jingyi Ju, Zhihao Li, and Yangang Wang. Reconstructing groups of people with hypergraph relational reasoning. In ICCV, pages 14873–14883, 2023. 1, 2, 3, 6, 7

    2023

  8. [16]

    Closely interactive human reconstruction with proxemics and physics-guided adaption

    Buzhen Huang, Chen Li, Chongyang Xu, Liang Pan, Yan- gang Wang, and Gim Hee Lee. Closely interactive human reconstruction with proxemics and physics-guided adaption. In CVPR, 2024. 2, 3, 5, 6, 7

    2024

  9. [17]

    Human3.6m: Large scale datasets and predic- tive methods for 3d human sensing in natural environments

    Catalin Ionescu, Dragos Papava, Vlad Olaru, and Cristian Sminchisescu. Human3.6m: Large scale datasets and predic- tive methods for 3d human sensing in natural environments. TPAMI, 36(7):1325–1339, 2014. 3

    2014

  10. [18]

    Coherent reconstruction of multiple humans from a single image

    Wen Jiang, Nikos Kolotouros, Georgios Pavlakos, Xiaowei Zhou, and Kostas Daniilidis. Coherent reconstruction of multiple humans from a single image. InCVPR, pages 5579– 5588, 2020. 1, 2, 3

    2020

  11. [19]

    Uv gaussians: Joint learning of mesh deformation and gaus- sian textures for human avatar modeling

    Yujiao Jiang, Qingmin Liao, Xiaoyu Li, Li Ma, Qi Zhang, Chaopeng Zhang, Zongqing Lu, and Ying Shan. Uv gaussians: Joint learning of mesh deformation and gaus- sian textures for human avatar modeling. arXiv preprint arXiv:2403.11589, 2024. 5

    2024 arXiv

  12. [20]

    Hifi4g: High-fidelity human performance rendering via compact gaussian splatting

    Yuheng Jiang, Zhehao Shen, Penghao Wang, Zhuo Su, Yu Hong, Yingliang Zhang, Jingyi Yu, and Lan Xu. Hifi4g: High-fidelity human performance rendering via compact gaussian splatting. In CVPR, 2024. 3

    2024

  13. [21]

    Exem- plar fine-tuning for 3d human model fitting towards in-the- wild 3d human pose estimation

    Hanbyul Joo, Natalia Neverova, and Andrea Vedaldi. Exem- plar fine-tuning for 3d human model fitting towards in-the- wild 3d human pose estimation. In 3DV, pages 42–52, 2021. 3

    2021

  14. [22]

    End-to-end recovery of human shape and pose

    Angjoo Kanazawa, Michael J Black, David W Jacobs, and Jitendra Malik. End-to-end recovery of human shape and pose. In CVPR, pages 7122–7131, 2018. 2, 3

    2018

  15. [23]

    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. TOG, 42(4):1–14, 2023. 2, 3

    2023

  16. [24]

    Oc- cluded human mesh recovery

    Rawal Khirodkar, Shashank Tripathi, and Kris Kitani. Oc- cluded human mesh recovery. In CVPR, pages 1715–1725,

  17. [25]

    Segment any- thing

    Alexander Kirillov, Eric Mintun, Nikhila Ravi, Hanzi Mao, Chloe Rolland, Laura Gustafson, Tete Xiao, Spencer White- head, Alexander C Berg, Wan-Yen Lo, et al. Segment any- thing. In ICCV, pages 4015–4026, 2023. 1, 2, 4, 5

    2023

  18. [26]

    Hugs: Human gaussian splats

    Muhammed Kocabas, Jen-Hao Rick Chang, James Gabriel, Oncel Tuzel, and Anurag Ranjan. Hugs: Human gaussian splats. In CVPR, 2024. 3

    2024

  19. [27]

    Unite the peo- ple: Closing the loop between 3d and 2d human representa- tions

    Christoph Lassner, Javier Romero, Martin Kiefel, Federica Bogo, Michael J Black, and Peter V Gehler. Unite the peo- ple: Closing the loop between 3d and 2d human representa- tions. In CVPR, pages 6050–6059, 2017. 3, 5

    2017

  20. [28]

    Guess the unseen: Dynamic 3d scene reconstruction from partial 2d glimpses

    Inhee Lee, Byungjun Kim, and Hanbyul Joo. Guess the unseen: Dynamic 3d scene reconstruction from partial 2d glimpses. In CVPR, 2024. 3, 8

    2024

  21. [29]

    Gart: Gaussian articulated template mod- els

    Jiahui Lei, Yufu Wang, Georgios Pavlakos, Lingjie Liu, and Kostas Daniilidis. Gart: Gaussian articulated template mod- els. In CVPR, 2024. 3

    2024

  22. [30]

    Coordinate transformer: Achieving single-stage multi- person mesh recovery from videos

    Haoyuan Li, Haoye Dong, Hanchao Jia, Dong Huang, Michael C Kampffmeyer, Liang Lin, and Xiaodan Liang. Coordinate transformer: Achieving single-stage multi- person mesh recovery from videos. In ICCV, pages 8744– 8753, 2023. 2

    2023

  23. [31]

    Cliff: Carrying location information in full frames into human pose and shape estimation

    Zhihao Li, Jianzhuang Liu, Zhensong Zhang, Songcen Xu, and Youliang Yan. Cliff: Carrying location information in full frames into human pose and shape estimation. In ECCV, pages 590–606, 2022. 1, 2, 3, 4, 6, 7

    2022

  24. [32]

    Ani- matable gaussians: Learning pose-dependent gaussian maps for high-fidelity human avatar modeling

    Zhe Li, Zerong Zheng, Lizhen Wang, and Yebin Liu. Ani- matable gaussians: Learning pose-dependent gaussian maps for high-fidelity human avatar modeling. In CVPR, 2024. 3

    2024

  25. [33]

    Intergen: Diffusion-based multi-human motion generation under complex interactions

    Han Liang, Wenqian Zhang, Wenxuan Li, Jingyi Yu, and Lan Xu. Intergen: Diffusion-based multi-human motion generation under complex interactions. arXiv preprint arXiv:2304.05684, 2023. 3, 4, 6

    2023 arXiv

  26. [34]

    One-stage 3d whole-body mesh recovery with component aware transformer

    Jing Lin, Ailing Zeng, Haoqian Wang, Lei Zhang, and Yu Li. One-stage 3d whole-body mesh recovery with component aware transformer. In CVPR, pages 21159–21168, 2023. 3

    2023

  27. [35]

    Vastgaussian: Vast 3d gaussians for large scene reconstruction

    Jiaqi Lin, Zhihao Li, Xiao Tang, Jianzhuang Liu, Shiyong Liu, Jiayue Liu, Yangdi Lu, Xiaofei Wu, Songcen Xu, You- liang Yan, et al. Vastgaussian: Vast 3d gaussians for large scene reconstruction. arXiv preprint arXiv:2402.17427 ,

    arXiv

  28. [36]

    Smpl: A skinned multi- person linear model

    Matthew Loper, Naureen Mahmood, Javier Romero, Gerard Pons-Moll, and Michael J Black. Smpl: A skinned multi- person linear model. TOG, 34(6):1–16, 2015. 3

    2015

  29. [37]

    Autotrackanything, 2024

    Roman Lyskov. Autotrackanything, 2024. 1, 5

    2024

  30. [38]

    Neuralannot: Neural annotator for 3d human mesh training sets

    Gyeongsik Moon, Hongsuk Choi, and Kyoung Mu Lee. Neuralannot: Neural annotator for 3d human mesh training sets. In CVPRW, pages 2298–2306, 2022. 3

    2022

  31. [39]

    Human gaussian splatting: Real-time rendering of animatable avatars

    Arthur Moreau, Jifei Song, Helisa Dhamo, Richard Shaw, Yiren Zhou, and Eduardo P ´erez-Pellitero. Human gaussian splatting: Real-time rendering of animatable avatars. In CVPR, 2024. 3

    2024

  32. [40]

    Generative proxemics: A prior for 3d social interaction from images

    Lea M ¨uller, Vickie Ye, Georgios Pavlakos, Michael Black, and Angjoo Kanazawa. Generative proxemics: A prior for 3d social interaction from images. In CVPR, 2024. 2, 3, 6, 7

    2024

  33. [41]

    Improved denoising diffusion probabilistic models

    Alexander Quinn Nichol and Prafulla Dhariwal. Improved denoising diffusion probabilistic models. In ICML, 2021. 4

    2021

  34. [42]

    Ash: Animatable gaussian splats for efficient and photoreal human rendering

    Haokai Pang, Heming Zhu, Adam Kortylewski, Christian Theobalt, and Marc Habermann. Ash: Animatable gaussian splats for efficient and photoreal human rendering. InCVPR,

  35. [43]

    Georgios Pavlakos, Vasileios Choutas, Nima Ghorbani, Timo Bolkart, Ahmed A. A. Osman, Dimitrios Tzionas, and Michael J. Black. Expressive body capture: 3D hands, face, and body from a single image. In CVPR, pages 10975– 10985, 2019. 3, 5

    2019

  36. [44]

    Human mesh recovery from multiple shots

    Georgios Pavlakos, Jitendra Malik, and Angjoo Kanazawa. Human mesh recovery from multiple shots. In CVPR, pages 1485–1495, 2022. 3

    2022

  37. [45]

    3dgs-avatar: Animatable avatars via deformable 3d gaussian splatting

    Zhiyin Qian, Shaofei Wang, Marko Mihajlovic, Andreas Geiger, and Siyu Tang. 3dgs-avatar: Animatable avatars via deformable 3d gaussian splatting. In CVPR, 2024. 3

    2024

  38. [46]

    Psvt: End-to-end multi-person 3d pose and shape estimation with progressive video transformers

    Zhongwei Qiu, Qiansheng Yang, Jian Wang, Haocheng Feng, Junyu Han, Errui Ding, Chang Xu, Dongmei Fu, and Jingdong Wang. Psvt: End-to-end multi-person 3d pose and shape estimation with progressive video transformers. In CVPR, pages 21254–21263, 2023. 2

    2023

  39. [47]

    U-net: Convolutional networks for biomedical image segmentation

    Olaf Ronneberger, Philipp Fischer, and Thomas Brox. U-net: Convolutional networks for biomedical image segmentation. In MICCAI, pages 234–241, 2015. 2

    2015

  40. [48]

    Physcap: Physically plausible monocular 3d motion capture in real time

    Soshi Shimada, Vladislav Golyanik, Weipeng Xu, and Chris- tian Theobalt. Physcap: Physically plausible monocular 3d motion capture in real time. ACM Transactions on Graphics (ToG), 39(6):1–16, 2020. 8

    2020

  41. [49]

    Score-guided diffusion for 3d human recovery

    Anastasis Stathopoulos, Ligong Han, and Dimitris Metaxas. Score-guided diffusion for 3d human recovery. In CVPR,

  42. [50]

    Pose priors from language models

    Sanjay Subramanian, Evonne Ng, Lea M ¨uller, Dan Klein, Shiry Ginosar, and Trevor Darrell. Pose priors from language models. arXiv preprint arXiv:2405.03689, 2024. 3

    2024 arXiv

  43. [51]

    Monocular, one-stage, regression of multiple 3d people

    Yu Sun, Qian Bao, Wu Liu, Yili Fu, Michael J Black, and Tao Mei. Monocular, one-stage, regression of multiple 3d people. In ICCV, pages 11179–11188, 2021. 2

    2021

  44. [52]

    Putting people in their place: Monocular regression of 3d people in depth

    Yu Sun, Wu Liu, Qian Bao, Yili Fu, Tao Mei, and Michael J Black. Putting people in their place: Monocular regression of 3d people in depth. In CVPR, pages 13243–13252, 2022. 1, 2, 3, 6, 7

    2022

  45. [53]

    Tiktok terms of service

    TikTok. Tiktok terms of service. https : / / www.tiktok.com/legal/page/row/terms- of- service/en, 2021. Accessed: 2024-11-01. 6

    2021

  46. [54]

    TikTok. Tiktok. https://www.tiktok.com/ , 2021. Accessed: 2024-11-01. 6

    2021

  47. [55]

    Capturing hands in action using discriminative salient points and physics sim- ulation

    Dimitrios Tzionas, Luca Ballan, Abhilash Srikantha, Pablo Aponte, Marc Pollefeys, and Juergen Gall. Capturing hands in action using discriminative salient points and physics sim- ulation. IJCV, 118(2):172–193, 2016. 5

    2016

  48. [56]

    Multiphys: Multi-person physics-aware 3d motion estimation

    Nicolas Ugrinovic, Boxiao Pan, Georgios Pavlakos, De- spoina Paschalidou, Bokui Shen, Jordi Sanchez-Riera, Francesc Moreno-Noguer, and Leonidas Guibas. Multiphys: Multi-person physics-aware 3d motion estimation. InCVPR,

  49. [57]

    Recovering accurate 3d human pose in the wild using imus and a moving camera

    Timo von Marcard, Roberto Henschel, Michael Black, Bodo Rosenhahn, and Gerard Pons-Moll. Recovering accurate 3d human pose in the wild using imus and a moving camera. In ECCV, 2018. 6

    2018

  50. [58]

    Outdoor markerless motion capture with sparse handheld video cameras

    Yangang Wang, Yebin Liu, Xin Tong, Qionghai Dai, and Ping Tan. Outdoor markerless motion capture with sparse handheld video cameras. TVCG, 24(5):1856–1866, 2017. 5

    2017

  51. [59]

    Image quality assessment: from error visibility to structural similarity

    Zhou Wang, Alan C Bovik, Hamid R Sheikh, and Eero P Simoncelli. Image quality assessment: from error visibility to structural similarity. TIP, pages 600–612, 2004. 5

    2004

  52. [60]

    Crowd3d: Towards hundreds of people reconstruction from a single image

    Hao Wen, Jing Huang, Huili Cui, Haozhe Lin, Yu-Kun Lai, Lu Fang, and Kun Li. Crowd3d: Towards hundreds of people reconstruction from a single image. In CVPR, pages 8937– 8946, 2023. 3

    2023

  53. [61]

    Gomavatar: Efficient an- imatable human modeling from monocular video using gaussians-on-mesh

    Jing Wen, Xiaoming Zhao, Zhongzheng Ren, Alexander G Schwing, and Shenlong Wang. Gomavatar: Efficient an- imatable human modeling from monocular video using gaussians-on-mesh. In CVPR, 2024. 3

    2024

  54. [62]

    Inter-x: Towards versatile human-human interaction analysis

    Liang Xu, Xintao Lv, Yichao Yan, Xin Jin, Shuwen Wu, Congsheng Xu, Yifan Liu, Yizhou Zhou, Fengyun Rao, Xingdong Sheng, Yunhui Liu, Wenjun Zeng, and Xiaokang Yang. Inter-x: Towards versatile human-human interaction analysis. arXiv preprint arXiv:2312.16051, 2023. 3, 6

    2023 arXiv

  55. [63]

    Vit- pose++: Vision transformer for generic body pose estima- tion

    Yufei Xu, Jing Zhang, Qiming Zhang, and Dacheng Tao. Vit- pose++: Vision transformer for generic body pose estima- tion. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2023. 1, 2, 4, 5, 6

    2023

  56. [64]

    Occgaussian: 3d gaus- sian splatting for occluded human rendering

    Jingrui Ye, Zongkai Zhang, Yujiao Jiang, Qingmin Liao, Wenming Yang, and Zongqing Lu. Occgaussian: 3d gaus- sian splatting for occluded human rendering. arXiv preprint arXiv:2404.08449, 2024. 3

    2024 arXiv

  57. [65]

    Hi4d: 4d instance segmen- tation of close human interaction

    Yifei Yin, Chen Guo, Manuel Kaufmann, Juan Jose Zarate, Jie Song, and Otmar Hilliges. Hi4d: 4d instance segmen- tation of close human interaction. In CVPR, pages 17016– 17027, 2023. 6

    2023

  58. [66]

    Monocular 3d pose and shape estimation of multiple people in natural scenes-the importance of multiple scene constraints

    Andrei Zanfir, Elisabeta Marinoiu, and Cristian Sminchis- escu. Monocular 3d pose and shape estimation of multiple people in natural scenes-the importance of multiple scene constraints. In CVPR, pages 2148–2157, 2018. 1, 2, 3

    2018

  59. [67]

    Deep network for the in- tegrated 3d sensing of multiple people in natural images

    Andrei Zanfir, Elisabeta Marinoiu, Mihai Zanfir, Alin-Ionut Popa, and Cristian Sminchisescu. Deep network for the in- tegrated 3d sensing of multiple people in natural images. NeurIPS, 31, 2018. 3

    2018

  60. [68]

    Body meshes as points

    Jianfeng Zhang, Dongdong Yu, Jun Hao Liew, Xuecheng Nie, and Jiashi Feng. Body meshes as points. In CVPR, pages 546–556, 2021. 3

    2021

  61. [69]

    The unreasonable effectiveness of deep features as a perceptual metric

    Richard Zhang, Phillip Isola, Alexei A Efros, Eli Shechtman, and Oliver Wang. The unreasonable effectiveness of deep features as a perceptual metric. In CVPR, pages 586–595,

  62. [70]

    Gps- gaussian: Generalizable pixel-wise 3d gaussian splatting for real-time human novel view synthesis

    Shunyuan Zheng, Boyao Zhou, Ruizhi Shao, Boning Liu, Shengping Zhang, Liqiang Nie, and Yebin Liu. Gps- gaussian: Generalizable pixel-wise 3d gaussian splatting for real-time human novel view synthesis. In CVPR, 2024. 3

    2024

  63. [71]

    On the continuity of rotation representations in neural networks

    Yi Zhou, Connelly Barnes, Jingwan Lu, Jimei Yang, and Hao Li. On the continuity of rotation representations in neural networks. In CVPR, pages 5745–5753, 2019. 3

    2019

Pith tools

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