Pith. sign in

REVIEW 3 major objections 5 minor 53 references

X as Supervision: Contending with Depth Ambiguity in Unsupervised Monocular 3D Pose Estimation

T0 review · 3 major / 5 minor · reviewed 2026-08-12 · deepseek-v4-flash

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

Pith's one-line read Unsupervised monocular 3D pose estimation should output several depth hypotheses, not one.

desk verdict Novel multi-hypothesis depth decoding with solid ablations, but the abstract's SOTA claim rests on an oracle best-of-N metric; the confidence-based number is more honest and still competitive. read the letter →

arxiv 2411.13026 v1 pith:VJAP6ZLL submitted 2024-11-20 cs.CV

Yuchen Yang , Xuanyi Liu , Xing Gao , Zhihang Zhong , Xiao Sun This is my paper · ORCID
classification cs.CV
keywords unsupervised3Dposeestimationdepthambiguitymulti-hypothesisdetectorSMPLmodelpretexttasksWinner-Takes-AlllossGCNdiscriminatorsyntheticrendering
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 claims that previous unsupervised methods for monocular 3D pose estimation have been formulated almost entirely in 2D space and therefore ignore the inherent depth ambiguity that makes the recovery of depth from a single image a multi-solution problem. The authors argue that the correct response is to let the network produce multiple depth hypotheses per joint, then use human-body priors to keep those hypotheses anatomically plausible. They build those priors from the SMPL parametric body model, sampling pose and shape parameters from hand-set truncated Gaussian distributions and enforcing plausibility through a GCN-based discriminator and rendered synthetic images. The reported result is state-of-the-art unsupervised performance: 72.2 mm MPJPE on Human3.6M and 11.50 cm on MPI-INF-3DHP, with the most confident hypothesis alone already beating all compared methods that rely on extra views, templates, or paired joints. The same recipe transfers to in-the-wild data and to dog pose estimation through the SMAL model.

What carries the argument

The multi-hypothesis detector aggregates the 3D heatmap into a marginal depth heatmap $H_{\mathrm{depth}}$, finds local peaks, selects the top $N_{\mathrm{hypo}}$ candidates, and refines each with windowed weighted averaging, producing $N_{\mathrm{hypo}}$ depth hypotheses per joint. It is trained with the Winner-Takes-All loss, which penalizes only the hypothesis closest to the target at each step and thereby preserves multimodal responses. Around this detector, the SMPL-driven pretext tasks sample pose and shape parameters from hand-set truncated Gaussian distributions, feed 3D keypoints and bones into a dual-representation GCN-based discriminator under an LSGAN objective, and render synthetic SMPL images with SURREAL textures for direct regression. The 2D anchor is provided by the Mask-as-Supervision reconstruction loss, and the whole stack is combined into a single weighted objective.

What would settle it

Train the full loss stack on Human3.6M with a vanilla single-hypothesis soft-argmax decoder while keeping all SMPL pretext tasks and hyperparameters; if its MPJPE stays within a few millimeters of the 72.2 mm multi-hypothesis best, the multi-hypothesis decoding is not the operative cause of the reported improvement.

Watch

Extended reading notes

Core claim

Depth ambiguity in unsupervised monocular 3D pose estimation is not measurement noise to be averaged away; it is an inherent multi-solution structure. The paper claims that a detector that decodes several depth peaks from the marginal depth heatmap, trained with a Winner-Takes-All loss, preserves this structure instead of suppressing plausible alternatives, and that SMPL-driven pretext tasks regularize the solution space toward plausible human structures. Concretely, the framework reports best-of-hypotheses MPJPE of 72.2 mm on Human3.6M and 11.50 cm on MPI-INF-3DHP, both below every compared unsupervised method, while the confident hypothesis alone already outperforms prior work at 93.7 mm on Human3.6M. Ablations show that removing the multi-hypothesis decoding costs about 20 mm, and that the dual-representation GCN discriminator outperforms an MLP discriminator by a wide margin. The authors further claim that the approach scales with unlabeled in-the-wild data and generalizes to animal pose estimation by swapping SMPL for SMAL.

Load-bearing premise

The load-bearing premise is that the hand-set truncated Gaussian ranges for each SMPL joint angle faithfully approximate the real distribution of natural human poses; if those ranges are unrepresentative, the discriminator and render loss will bias predictions away from plausible structures.

Editorial extensions

If this is right

  • Unsupervised monocular 3D pose estimation can be built without multi-view images, reference images, hand-designed templates, or paired joints; the paper's best results use none of these and still outperform methods that do.
  • Multi-hypothesis decoding changes only the heatmap decoding step and can be written as matrix operations, so handling depth ambiguity adds no extra network parameters or inference overhead.
  • The SMPL prior transfers across species: replacing SMPL with SMAL yields plausible unsupervised 3D dog poses on the StanfordExtra dataset without animal-specific distribution curation.
  • The method scales with unlabeled in-the-wild data: progressively adding Human3.6M, MPI-INF-3DHP, and TikTok training data improves MPII evaluation performance beyond the prior state of the art at every stage.
  • The gap between the confident hypothesis and the best hypothesis is a quantifiable measure of remaining depth ambiguity, indicating that the detector genuinely entertains multiple depths rather than collapsing to a single mode.

Reading between the lines

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

  • If the multi-solution framing is right, a natural testable extension is to replace the hand-set truncated Gaussians with distributions fitted to large motion-capture collections; this should tighten the plausibility constraint and likely improve the confident-hypothesis error while narrowing the conf-to-best gap.
  • The peak-selection-plus-WTA recipe is generic and could be applied to other depth-ambiguous regression tasks, such as single-view object shape reconstruction, wherever a parametric structure prior is available.
  • The pipeline still requires pre-extracted binary human masks as the 2D anchor, so 'unsupervised' here assumes a segmentation signal; replacing these masks with an automatic segmenter would be the step that makes the full pipeline label-free in practice.
  • Because the paper reports best-of-hypotheses error, a practical deployment would need a confidence-calibrated rule for choosing among the emitted hypotheses; the paper does not propose one, and the confident hypothesis is measurably worse than the best.
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 an unsupervised monocular 3D pose estimation framework, X as Supervision, whose core idea is to treat depth ambiguity as a multi-solution problem. The method introduces a multi-hypothesis detector that decodes several depth hypotheses from local peaks of a depth marginal heatmap and trains them with a Winner-Takes-All loss. Supervision comes from two SMPL-based pretext tasks: a GCN discriminator on keypoint and bone graphs, and a regression loss on rendered synthetic SMPL images, combined with an existing 2D mask reconstruction loss. Experiments on Human3.6M and MPI-INF-3DHP report MPJPE, with ablations, data scale-up experiments, and a qualitative animal-pose extension. The paper claims state-of-the-art unsupervised 3D pose estimation performance.

Significance. The framing of depth ambiguity as an explicit multi-solution problem in unsupervised monocular 3D pose estimation is timely and, to my knowledge, not directly addressed by prior work. The multi-hypothesis detector is lightweight, adding no network parameters, and the ablations show consistent gains from both the multi-hypothesis detector and the SMPL-based constraints. The SURREAL variant demonstrates that the framework is not wholly dependent on the hand-defined synthetic prior. The paper is also honest in reporting both oracle and confidence-based columns. However, the headline state-of-the-art claim rests on an oracle best-of-N evaluation that is not achievable at test time, and the confidence-based numbers are substantially worse; this gap must be resolved before the central claim can be accepted.

major comments (3)
  1. [Section 4.2.1/4.2.2, Table 1] The abstract and introduction state that the approach demonstrates state-of-the-art unsupervised 3D pose estimation, but the only Human3.6M result supporting this is Ours-SurMH (best) (72.2 mm MPJPE), which is obtained by selecting, at evaluation time, the hypothesis closest to ground truth. The deployable Ours-SurMH (conf) result is 93.7 mm, which is 21.5 mm worse and is not clearly state-of-the-art relative to several methods listed in the same table (e.g., Honari et al. [13], 73.8 mm, albeit using additional supervision). I request that the confidence-based numbers be used as the headline result, or that a test-time selection rule be provided and evaluated; the best-of-N column should be explicitly labeled as an oracle upper bound.
  2. [Section 4.2.1] The confidence-based selection rule is defined as choosing the hypothesis with the maximum peak response, but no evidence is provided that peak response is a reliable predictor of correctness. The 21.5 mm gap between 'conf' and 'best' in Table 1 indicates that the current rule leaves most of the potential benefit of multiple hypotheses unrealized. Please add an analysis of confidence quality, such as coverage of the best hypothesis among the top-k by confidence, or a learned selection module.
  3. [Section 3.3.2 and Appendix C.1] The synthetic prior is constructed from hand-defined truncated Gaussian ranges for each SMPL joint (e.g., knee and hip ranges [45,60], [10,10] degrees), without validation against an empirical distribution of natural human poses. Since the GAN and render losses are the only direct 3D constraints, unrepresentative ranges could bias predictions. The SURREAL variant in Table 4 (Sur-Texture, 88.56 mm vs Syn-Texture, 105.61 mm) shows that prior quality matters and mitigates the concern, but the paper should either validate the chosen ranges against motion-capture statistics or relegate the Syn setting to an ablation and use a data-driven prior for the main claims.
minor comments (5)
  1. [Section 3.3.1] The paper calls the human binary mask Mgt 'easily attainable,' but a segmentation mask is still a form of annotation; please clarify the exact source of masks in training and how this fits the claimed fully unsupervised setting.
  2. [Table 1 vs Table 2] MPJPE is reported in millimeters in Table 1 and centimeters in Table 2 without a unit label in each table header; please standardize the units or add explicit unit labels to avoid confusion.
  3. [Section 4.4.2] The animal pose estimation experiment is qualitative only; please add quantitative metrics or a comparison to a simple baseline to support the generalization claim.
  4. [Footnote 1 and Acknowledgments] The Wikipedia citation for multimodal distribution is informal for a journal paper, and 'Corresponding arthor' in the author footnote is a typo; both should be corrected.
  5. [Figure 4 and Table 4] The correspondence between the appearance settings in Figure 4 and the rows of Table 4 is not explicitly marked; adding labels such as 'Syn-Seg', 'Syn-Texture', and 'Sur-Texture' directly in the figure would improve readability.

Circularity Check

1 steps flagged · score 6.0 of 10

The headline SOTA number is an oracle best-of-N selection: 72.2 mm MPJPE is the hypothesis closest to ground truth, while the deployable confidence-based result is 93.7 mm.

  1. other [Abstract; Sec. 4.2.1 (Baselines); Table 1]
    "Consequently, our approach demonstrates state-of-the-art unsupervised 3D pose estimation performance on various human datasets. ... best utilizes the hypothesis closest to the ground truth, performing the upper bound for addressing depth ambiguity. ... Ours-SurMH (best) ... 72.2"

    The reported SOTA result is not a committed model prediction: 'best' is defined as the hypothesis closest to the ground truth at test time, so the 72.2 mm row is by construction the minimum MPJPE over the N_hypo=3 hypotheses after seeing the label. The deployable confidence-selected row is 93.7 mm, 21.5 mm worse, and no longer beats prior methods such as Honari et al. [13] (73.8 mm). The abstract presents this oracle-based value as unqualified 'state-of-the-art' performance, so the central advertised prediction reduces to a test-time ground-truth selection rule. The paper discloses the upper-bound nature in Sec. 4.2.1, but the headline claim still rests on the oracle metric.

full rationale

Most of the derivation chain is self-contained and not circular. The model is trained with 2D mask reconstruction, a GCN discriminator over SMPL-synthesized skeletons, and render losses on synthetic SMPL images; evaluation on Human3.6M and MPI-INF-3DHP uses held-out labels that never enter training. The hand-set truncated Gaussian pose priors are a validity assumption rather than a circular dependency, and the SURREAL ablation shows the framework does not depend on those exact ranges. The self-citation to Mask as Supervision [49] supplies a component loss and is not the source of the depth-ambiguity claim. The one concrete circular element is in the headline metric: 'best' selects the hypothesis closest to ground truth, so the 72.2 mm SOTA value is an oracle upper bound, while the deployable confidence-selected result (93.7 mm) is 21.5 mm worse and not state-of-the-art. Because the abstract presents the oracle value as 'state-of-the-art' without qualification, a central advertised prediction reduces by construction to test-time ground-truth selection. This is partial circularity in the reported claim, not in the training objective itself.

Assumptions & free parameters 7 free parameters · 5 assumptions · 0 invented entities

The central claim rests on a parametric body model (SMPL), manually specified sampling priors, and a mask supervision channel. While no new physical entities are introduced, the hand-tuned priors and loss weights constitute free parameters that shape the learned solution. The method does not derive from first principles; it transfers human-structure knowledge from SMPL into the detector.

free parameters (7)
  • Number of hypotheses N_hypo = 3
    Chosen empirically; Appendix C.1 states N_hypo=3. Affects the capacity of the multi-solution detector and the WTA loss.
  • Local window size N_w = 15
    Appendix C.1 sets N_w=15 for depth heatmap of size 18x64. Controls depth hypothesis refinement.
  • Loss weights lambda_m, lambda_g, lambda_r = 2e-2, 1.0, 0.5
    Set to keep loss scales equal; Appendix C.1.
  • SMPL pose truncated Gaussian ranges per joint = e.g., knee/hip [45,60] deg, shoulder/elbow [90,90] deg, spine [60,20] deg, global rotation [-5,180,-5]
    Hand-defined in Appendix C.1 to generate plausible synthetic poses. These are the key prior that shapes the discriminator and render supervision.
  • SMPL shape truncation width = [-1.5, 1.5]
    Appendix C.1. Controls body shape diversity in synthetic set.
  • T-pose probability = 0.4
    Appendix C.1. Each pose parameter has 40% chance to stay at T-pose to anchor the prior.
  • Number of GraphConv modules in discriminator = 2
    Appendix C.1. Architecture choice for the GCN discriminator.
assumptions (5)
  • domain assumption SMPL model provides a valid parametric representation of human body pose and shape.
    Section 3.3.1. The mesh V = SMPL(theta, beta, xi) is treated as the source of ground-truth-like 3D keypoints for synthetic data.
  • ad hoc to paper The hand-defined truncated Gaussian distributions approximate the empirical distribution of natural human poses.
    Appendix C.1. The joint angle ranges are chosen by the authors based on anatomical intuition, not learned from data; if they are unrepresentative, the discriminator and render losses will bias predictions.
  • domain assumption Binary human masks M_gt are available for training images without 3D annotations.
    Section 3.3.1. The mask loss requires ground-truth 2D masks; obtaining masks may need segmentation or manual annotation, which weakens the 'unsupervised' claim.
  • ad hoc to paper The depth heatmap of a joint is multimodal and its local peaks correspond to plausible depth hypotheses.
    Section 3.2, Equations (3)-(5). The multi-hypothesis detector relies on top-K peak selection; if the heatmap is unimodal or poorly calibrated, hypotheses collapse and WTA cannot recover multiple solutions.
  • domain assumption The LSGAN discriminator with Winner-Takes-All provides a stable and useful training signal.
    Section 3.3.2, Equations (14)-(15). Standard GAN training dynamics are assumed to converge without mode collapse or oscillation.

how reviews work

0 comments
Cite this review

Pith. "Pith review of X as Supervision: Contending with Depth Ambiguity in Unsupervised Monocular 3D Pose Estimation." pith.science (2026). https://pith.science/paper/VJAP6ZLL

@misc{pith2026241113026,
  author       = {Pith},
  title        = {Pith review of: X as Supervision: Contending with Depth Ambiguity in Unsupervised Monocular 3D Pose Estimation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/VJAP6ZLL}},
  note         = {Machine review of arXiv:2411.13026}
}
read the original abstract

Recent unsupervised methods for monocular 3D pose estimation have endeavored to reduce dependence on limited annotated 3D data, but most are solely formulated in 2D space, overlooking the inherent depth ambiguity issue. Due to the information loss in 3D-to-2D projection, multiple potential depths may exist, yet only some of them are plausible in human structure. To tackle depth ambiguity, we propose a novel unsupervised framework featuring a multi-hypothesis detector and multiple tailored pretext tasks. The detector extracts multiple hypotheses from a heatmap within a local window, effectively managing the multi-solution problem. Furthermore, the pretext tasks harness 3D human priors from the SMPL model to regularize the solution space of pose estimation, aligning it with the empirical distribution of 3D human structures. This regularization is partially achieved through a GCN-based discriminator within the discriminative learning, and is further complemented with synthetic images through rendering, ensuring plausible estimations. Consequently, our approach demonstrates state-of-the-art unsupervised 3D pose estimation performance on various human datasets. Further evaluations on data scale-up and one animal dataset highlight its generalization capabilities. Code will be available at https://github.com/Charrrrrlie/X-as-Supervision.

Figures

Figures reproduced from arXiv: 2411.13026 by the authors.

Figure 1
Figure 1. Challenges. (a) illustrates the pretext tasks of related methods, which localize keypoints by various structure representations for [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Framework overview of the X as Supervision. It contains a multi-hypothesis detector and novel pretext tasks for unsupervised [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Visualization of the predicted multiple hypotheses on [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figures from the paper (5 more)
Figure 4
Figure 4. Figure 4: Illustration of different tex￾tures. From top to down: SURREAL [45] human texture, part segmentation of SMPL [25] and SMAL [53]. 3.3.3. Loss According to Eq. (8), Eq. (14) and Eq. (16), the final objec￾tive for the detector is defined as a weighted combination of the p…
Figure 5
Figure 5. Figure 5: Performance on in-the-wild MPII dataset. Each node [PITH_FULL_IMAGE:figures/full_fig_p008_5.png]
Figure 7
Figure 7. Figure 7: Visualization of the predicted 2D projected poses and [PITH_FULL_IMAGE:figures/full_fig_p008_7.png]
Figure 8
Figure 8. Figure 8: Network structure of GCN-based discriminator. [PITH_FULL_IMAGE:figures/full_fig_p012_8.png]
Figure 9
Figure 9. Figure 9: Qualitative results on Human3.6M (top) and MPI-INF-3DHP (bottom). From left to right in each subfigure: predictions on 2D, [PITH_FULL_IMAGE:figures/full_fig_p013_9.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

53 extracted references · 44 canonical work pages

  1. [13]

    Unsupervised 3d keypoint es- timation with multi-view geometry

    Sina Honari and Pascal Fua. Unsupervised 3d keypoint es- timation with multi-view geometry. In 3D Vision, 2024 the 11th International Conference on. IEEE, 2024. 1, 2, 3, 6, 7

    work page 2024

  2. [1]

    2d human pose estimation: New benchmark and state of the art analysis

    Mykhaylo Andriluka, Leonid Pishchulin, Peter Gehler, and Bernt Schiele. 2d human pose estimation: New benchmark and state of the art analysis. In IEEE Conference on Com- puter Vision and Pattern Recognition, 2014. 6

    work page 2014

  3. [2]

    Who left the dogs out?: 3D animal reconstruction with expectation maximization in the loop

    Benjamin Biggs, Oliver Boyne, James Charles, Andrew Fitzgibbon, and Roberto Cipolla. Who left the dogs out?: 3D animal reconstruction with expectation maximization in the loop. In ECCV, 2020. 2, 6

    work page 2020

  4. [3]

    Within the dynamic con- text: Inertia-aware 3d human modeling with pose sequence

    Yutong Chen, Yifan Zhan, Zhihang Zhong, Wei Wang, Xiao Sun, Yu Qiao, and Yinqiang Zheng. Within the dynamic con- text: Inertia-aware 3d human modeling with pose sequence. arXiv preprint arXiv:2403.19160, 2024. 1

    arXiv 2024

  5. [4]

    Expressive whole-body con- trol for humanoid robots

    Xuxin Cheng, Yandong Ji, Junming Chen, Ruihan Yang, Ge Yang, and Xiaolong Wang. Expressive whole-body con- trol for humanoid robots. arXiv preprint arXiv:2402.16796,

    arXiv

  6. [5]

    Imagenet: A large-scale hierarchical image database

    Jia Deng, Wei Dong, Richard Socher, Li-Jia Li, Kai Li, and Li Fei-Fei. Imagenet: A large-scale hierarchical image database. In 2009 IEEE conference on computer vision and pattern recognition, pages 248–255. Ieee, 2009. 1

    2009

  7. [6]

    Unsupervised learning of disentangled representations from video

    Emily L Denton et al. Unsupervised learning of disentangled representations from video. Advances in Neural Information Processing Systems, 30, 2017. 7

    work page 2017

  8. [7]

    3d human reconstruction in the wild with synthetic data using generative models

    Yongtao Ge, Wenjia Wang, Yongfan Chen, Hao Chen, and Chunhua Shen. 3d human reconstruction in the wild with synthetic data using generative models. arXiv preprint arXiv:2403.11111, 2024. 5

    arXiv 2024

Show all 53 references
  1. [8]

    Multiple choice learning: Learning to produce multiple structured outputs

    Abner Guzm ´an-rivera, Dhruv Batra, and Pushmeet Kohli. Multiple choice learning: Learning to produce multiple structured outputs. In Advances in Neural Information Pro- cessing Systems. Curran Associates, Inc., 2012. 2

    2012

  2. [9]

    Inductive representation learning on large graphs

    Will Hamilton, Zhitao Ying, and Jure Leskovec. Inductive representation learning on large graphs. Advances in neural information processing systems, 30, 2017. 5

    2017

  3. [10]

    Multiple view ge- ometry in computer vision

    Richard Hartley and Andrew Zisserman. Multiple view ge- ometry in computer vision . Cambridge University Press,

  4. [11]

    Deep residual learning for image recognition

    Kaiming He, Xiangyu Zhang, Shaoqing Ren, and Jian Sun. Deep residual learning for image recognition. In Proceed- ings of the IEEE conference on computer vision and pattern recognition, pages 770–778, 2016. 1

    2016

  5. [12]

    Autolink: Self-supervised learning of human skeletons and object out- lines by linking keypoints

    Xingzhe He, Bastian Wandt, and Helge Rhodin. Autolink: Self-supervised learning of human skeletons and object out- lines by linking keypoints. Advances in Neural Information Processing Systems, 35:36123–36141, 2022. 1, 2, 3

    2022

  6. [14]

    Temporal representation learning on monocular videos for 3d human pose estimation

    Sina Honari, Victor Constantin, Helge Rhodin, Mathieu Salzmann, and Pascal Fua. Temporal representation learning on monocular videos for 3d human pose estimation. IEEE Transactions on Pattern Analysis and Machine Intelligence,

  7. [15]

    Catalin Ionescu, Dragos Papava, Vlad Olaru, and Cristian Sminchisescu. Human3. 6m: Large scale datasets and pre- dictive methods for 3d human sensing in natural environ- ments. IEEE Transactions on Pattern Analysis and Machine Intelligence, 36(7):1325–1339, 2013. 1, 2, 4, 6

    2013

  8. [16]

    Learning high fi- delity depths of dressed humans by watching social media dance videos

    Yasamin Jafarian and Hyun Soo Park. Learning high fi- delity depths of dressed humans by watching social media dance videos. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 12753– 12762, 2021. 6

    2021

  9. [17]

    Unsupervised learning of object landmarks through conditional image generation

    Tomas Jakab, Ankush Gupta, Hakan Bilen, and Andrea Vedaldi. Unsupervised learning of object landmarks through conditional image generation. Advances in neural informa- tion processing systems, 31, 2018. 2

    2018

  10. [18]

    Adam: A method for stochastic opti- mization

    Diederik P Kingma. Adam: A method for stochastic opti- mization. arXiv preprint arXiv:1412.6980, 2014. 1

    2014 arXiv

  11. [19]

    Semi-supervised classi- fication with graph convolutional networks

    Thomas N Kipf and Max Welling. Semi-supervised classi- fication with graph convolutional networks. arXiv preprint arXiv:1609.02907, 2016. 2

    2016 arXiv

  12. [20]

    Self-supervised 3d human pose estimation via part guided novel image synthesis

    Jogendra Nath Kundu, Siddharth Seth, Varun Jampani, Mu- galodi Rakesh, R Venkatesh Babu, and Anirban Chakraborty. Self-supervised 3d human pose estimation via part guided novel image synthesis. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognitio...

    2020

  13. [21]

    Kinematic-structure-preserved representation for unsupervised 3d human pose estimation

    Jogendra Nath Kundu, Siddharth Seth, MV Rahul, Muga- lodi Rakesh, Venkatesh Babu Radhakrishnan, and Anirban Chakraborty. Kinematic-structure-preserved representation for unsupervised 3d human pose estimation. In Proceed- ings of the AAAI Conference on Artificial Intelligence, ...

    2020

  14. [22]

    Stochastic multiple choice learning for training diverse deep ensembles

    Stefan Lee, Senthil Purushwalkam Shiva Prakash, Michael Cogswell, Viresh Ranjan, David Crandall, and Dhruv Batra. Stochastic multiple choice learning for training diverse deep ensembles. Advances in Neural Information Processing Sys- tems, 29, 2016. 2, 4

    2016

  15. [23]

    Human pose regression with residual log-likelihood estimation

    Jiefeng Li, Siyuan Bian, Ailing Zeng, Can Wang, Bo Pang, Wentao Liu, and Cewu Lu. Human pose regression with residual log-likelihood estimation. In Proceedings of the IEEE/CVF international conference on computer vision , pages 11025–11034, 2021. 4

    2021

  16. [24]

    Mhformer: Multi-hypothesis transformer for 3d human pose estimation

    Wenhao Li, Hong Liu, Hao Tang, Pichao Wang, and Luc Van Gool. Mhformer: Multi-hypothesis transformer for 3d human pose estimation. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages 13147–13156, 2022. 4

    2022

  17. [25]

    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. In Seminal Graphics Papers: Pushing the Boundaries, Volume 2, pages 851–866. 2023. 2, 5, 6

    2023

  18. [26]

    Perpetual humanoid control for real-time simulated avatars

    Zhengyi Luo, Jinkun Cao, Kris Kitani, Weipeng Xu, et al. Perpetual humanoid control for real-time simulated avatars. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 10895–10904, 2023. 1

    2023

  19. [27]

    Least squares genera- tive adversarial networks

    Xudong Mao, Qing Li, Haoran Xie, Raymond YK Lau, Zhen Wang, and Stephen Paul Smolley. Least squares genera- tive adversarial networks. In Proceedings of the IEEE inter- national conference on computer vision , pages 2794–2802,

  20. [28]

    Monocular 3d human pose estimation in the wild using improved cnn supervision

    Dushyant Mehta, Helge Rhodin, Dan Casas, Pascal Fua, Oleksandr Sotnychenko, Weipeng Xu, and Christian Theobalt. Monocular 3d human pose estimation in the wild using improved cnn supervision. In 3D Vision, 2017 Fifth International Conference on. IEEE, 2017. 1, 2, 4, 6, 7

    2017

  21. [29]

    Agora: Avatars in geography optimized for regression analysis

    Priyanka Patel, Chun-Hao P Huang, Joachim Tesch, David T Hoffmann, Shashank Tripathi, and Michael J Black. Agora: Avatars in geography optimized for regression analysis. In Proceedings of the IEEE/CVF Conference on Computer Vi- sion and Pattern Recognition, pages 13468–13478, 2021. 5

    2021

  22. [30]

    Learning to estimate 3d human pose and shape from a single color image

    Georgios Pavlakos, Luyang Zhu, Xiaowei Zhou, and Kostas Daniilidis. Learning to estimate 3d human pose and shape from a single color image. In Proceedings of the IEEE con- ference on computer vision and pattern recognition , pages 459–468, 2018. 5

    2018

  23. [31]

    Improving 2d human pose estimation in rare camera views with synthetic data

    Miroslav Purkrabek and Jiri Matas. Improving 2d human pose estimation in rare camera views with synthetic data. In 2024 IEEE 18th International Conference on Automatic Face and Gesture Recognition (FG), pages 1–9, 2024. 5

    2024

  24. [32]

    Neural scene decomposi- tion for multi-person motion capture

    Helge Rhodin, Victor Constantin, Isinsu Katircioglu, Math- ieu Salzmann, and Pascal Fua. Neural scene decomposi- tion for multi-person motion capture. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 7703–7713, 2019. 7

    2019

  25. [33]

    Barc: Learning to regress 3d dog shape from images by exploiting breed information

    Nadine Rueegg, Silvia Zuffi, Konrad Schindler, and Michael J Black. Barc: Learning to regress 3d dog shape from images by exploiting breed information. In Proceed- ings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 3876–3884, 2022. 8

    2022

  26. [34]

    Unsu- pervised human pose estimation through transforming shape templates

    Luca Schmidtke, Athanasios Vlontzos, Simon Ellershaw, Anna Lukens, Tomoki Arichi, and Bernhard Kainz. Unsu- pervised human pose estimation through transforming shape templates. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages 2484– 2...

    2021

  27. [35]

    Syn- thetic training for accurate 3d human pose and shape esti- mation in the wild

    Akash Sengupta, Ignas Budvytis, and Roberto Cipolla. Syn- thetic training for accurate 3d human pose and shape esti- mation in the wild. In British Machine Vision Conference (BMVC), 2020. 5

    2020

  28. [36]

    Ntu rgb+ d: A large scale dataset for 3d human activity anal- ysis

    Amir Shahroudy, Jun Liu, Tian-Tsong Ng, and Gang Wang. Ntu rgb+ d: A large scale dataset for 3d human activity anal- ysis. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 1010–1019, 2016. 1

    2016

  29. [37]

    Skeleton-based action recognition with directed graph neu- ral networks

    Lei Shi, Yifan Zhang, Jian Cheng, and Hanqing Lu. Skeleton-based action recognition with directed graph neu- ral networks. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 7912–7921,

  30. [38]

    Two- stream adaptive graph convolutional networks for skeleton- based action recognition

    Lei Shi, Yifan Zhang, Jian Cheng, and Hanqing Lu. Two- stream adaptive graph convolutional networks for skeleton- based action recognition. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages 12026–12035, 2019. 5

    2019

  31. [39]

    Self-supervised 3d human pose estimation from a single image

    Jose Sosa and David Hogg. Self-supervised 3d human pose estimation from a single image. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 4787–4796, 2023. 3, 6, 7

    2023

  32. [40]

    Selfpose3d: Self-supervised multi-person multi-view 3d pose estimation

    Vinkle Srivastav, Keqi Chen, and Nicolas Padoy. Selfpose3d: Self-supervised multi-person multi-view 3d pose estimation. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 2502–2512, 2024. 2

    2024

  33. [41]

    Bkind-3d: Self-supervised 3d keypoint discovery from multi-view videos

    Jennifer J Sun, Lili Karashchuk, Amil Dravid, Serim Ryou, Sonia Fereidooni, John C Tuthill, Aggelos Katsaggelos, Bingni W Brunton, Georgia Gkioxari, Ann Kennedy, et al. Bkind-3d: Self-supervised 3d keypoint discovery from multi-view videos. In Proceedings of the IEEE/CVF Con- ...

    2023

  34. [42]

    Integral human pose regression

    Xiao Sun, Bin Xiao, Fangyin Wei, Shuang Liang, and Yichen Wei. Integral human pose regression. In Proceedings of the European Conference on Computer Vision, pages 529–545,

  35. [43]

    Discovery of latent 3d key- points via end-to-end geometric reasoning.Advances in Neu- ral Information Processing Systems, 31, 2018

    Supasorn Suwajanakorn, Noah Snavely, Jonathan J Tomp- son, and Mohammad Norouzi. Discovery of latent 3d key- points via end-to-end geometric reasoning.Advances in Neu- ral Information Processing Systems, 31, 2018. 2, 6

    2018

  36. [44]

    Visualizing data using t-sne

    Laurens Van der Maaten and Geoffrey Hinton. Visualizing data using t-sne. Journal of machine learning research , 9 (11), 2008. 2

    2008

  37. [45]

    Learning from synthetic humans

    Gul Varol, Javier Romero, Xavier Martin, Naureen Mah- mood, Michael J Black, Ivan Laptev, and Cordelia Schmid. Learning from synthetic humans. In Proceedings of the IEEE conference on computer vision and pattern recogni- tion, pages 109–117, 2017. 5, 6, 1

    2017

  38. [46]

    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 European Conference on Computer Vision (ECCV), 2018. 1

    2018

  39. [47]

    Canonpose: Self-supervised monocu- lar 3d human pose estimation in the wild

    Bastian Wandt, Marco Rudolph, Petrissa Zell, Helge Rhodin, and Bodo Rosenhahn. Canonpose: Self-supervised monocu- lar 3d human pose estimation in the wild. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 13294–13304, 2021. 2

    2021

  40. [48]

    Spatial tempo- ral graph convolutional networks for skeleton-based action recognition

    Sijie Yan, Yuanjun Xiong, and Dahua Lin. Spatial tempo- ral graph convolutional networks for skeleton-based action recognition. In Proceedings of the AAAI conference on arti- ficial intelligence, 2018. 1

    2018

  41. [49]

    Mask as supervi- sion: Leveraging unified mask information for unsupervised 3d pose estimation

    Yuchen Yang, Yu Qiao, and Xiao Sun. Mask as supervi- sion: Leveraging unified mask information for unsupervised 3d pose estimation. In European Conference on Computer Vision, pages 38–55. Springer, 2024. 1, 2, 3, 4, 6, 7, 8

    2024

  42. [50]

    Synbody: Synthetic dataset with layered human models for 3d human perception and modeling

    Zhitao Yang, Zhongang Cai, Haiyi Mei, Shuai Liu, Zhaoxi Chen, Weiye Xiao, Yukun Wei, Zhongfei Qing, Chen Wei, Bo Dai, et al. Synbody: Synthetic dataset with layered human models for 3d human perception and modeling. In Proceedings of the IEEE/CVF International Conference on Co...

    2023

  43. [51]

    Towards alleviating the mod- eling ambiguity of unsupervised monocular 3d human pose estimation

    Zhenbo Yu, Bingbing Ni, Jingwei Xu, Junjie Wang, Cheng- long Zhao, and Wenjun Zhang. Towards alleviating the mod- eling ambiguity of unsupervised monocular 3d human pose estimation. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 8651–8660, 2021. 2

    2021

  44. [52]

    Unsupervised discovery of object land- marks as structural representations

    Yuting Zhang, Yijie Guo, Yixin Jin, Yijun Luo, Zhiyuan He, and Honglak Lee. Unsupervised discovery of object land- marks as structural representations. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recogni- tion, pages 2694–2703, 2018. 6

    2018

  45. [53]

    Silvia Zuffi, Angjoo Kanazawa, David Jacobs, and Michael J. Black. 3D menagerie: Modeling the 3D shape and pose of animals. In IEEE Conf. on Computer Vision and Pattern Recognition (CVPR), 2017. 6, 8 X as Supervision: Contending with Depth Ambiguity in Unsupervised Monocular 3...

    2017

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.