Pith. sign in

REVIEW 3 major objections 3 minor 46 references

HyperDiff: Hypergraph Guided Diffusion Model for 3D Human Pose Estimation

T0 review · 3 major / 3 minor · reviewed 2026-08-05 · deepseek-v4-flash

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

Pith's one-line read HyperDiff combines a conditional diffusion model with a multi-granularity hypergraph denoiser to lift 2D poses to 3D, and reports top accuracy on Human3.6M and MPI-INF-3DHP.

desk verdict Plausible new combination of diffusion and hypergraph denoising for 3D HPE, but the SOTA claim is unverifiable from the abstract alone; needs the experiments and protocol details. read the letter →

arxiv 2508.14431 v1 pith:U7ROJ6XB submitted 2025-08-20 cs.CV

Bing Han , Yuhua Huang , Pan Gao This is my paper · ORCID
classification cs.CV
keywords 3Dhumanposeestimation2D-to-3Dliftingdiffusionmodelhypergraphconvolutionmulti-granularitydepthambiguityHuman3.6MMPI-INF-3DHP
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

The paper introduces HyperDiff, a 2D-to-3D human pose lifting method that pairs a diffusion model with a hypergraph-based denoiser. It argues that diffusion naturally handles the uncertainty behind depth ambiguity and occlusion, while a multi-granularity hypergraph denoiser captures high-order correlations among joints that ordinary pairwise skeleton graphs miss. The authors report that this combination reaches top accuracy on Human3.6M and MPI-INF-3DHP, and that the number of diffusion steps can be adjusted to trade accuracy against computational cost. If the claim holds, modeling multi-joint dependencies beyond pairwise links is a practical lever for pose accuracy, not just a theoretical nicety.

What carries the argument

Conditional diffusion model: a forward process adds Gaussian noise to the 3D pose and a learned reverse process denoises it, conditioned on the 2D detections, producing a distribution over plausible poses rather than a single point estimate. Multi-granularity HyperGCN: a hypergraph convolution network whose hyperedges connect more than two joints at several scales, serving as the denoiser so that high-order joint correlations are available at every denoising step. The interaction of the two—diffusion absorbs depth and occlusion uncertainty while the hypergraph denoiser enforces structurally consistent poses—is the mechanism the paper credits for the accuracy gains.

What would settle it

Re-run Human3.6M with HyperDiff and the compared methods sharing identical 2D detections from one detector, under the same joint protocol and without refinement. If HyperDiff's mean per-joint position error is no longer lower, the state-of-the-art claim fails. Separately, replace the multi-granularity hypergraph denoiser with a plain graph denoiser of equal size under the same diffusion schedule; if accuracy holds, the hypergraph mechanism is not what carries the result.

Watch

Extended reading notes

Core claim

On the paper's own terms, the central claim is that the ambiguous 2D-to-3D lifting step can be treated as a conditional denoising problem and solved with a hypergraph-structured diffusion model. Given 2D keypoints, HyperDiff corrupts a 3D pose with noise and learns to remove it, so depth ambiguity and occlusion are handled as uncertainty rather than deterministic errors. The denoiser is a HyperGCN that builds hyperedges over joints at multiple granularities—pairs, limb groups, and larger joint sets—allowing it to reason about correlations a standard adjacency graph cannot express. The paper reports that this design achieves state-of-the-art accuracy on Human3.6M and MPI-INF-3DHP, and that th

Load-bearing premise

The headline accuracy claim depends on comparing HyperDiff with published methods under the same 2D keypoint detections and benchmark protocol; if the comparison used a stronger detector or a different protocol, the measured improvement would not be caused by HyperDiff's design.

Editorial extensions

If this is right

  • On standard lifting benchmarks, HyperDiff would displace the best previously published 2D-to-3D pose methods if the reported evaluation is replicated.
  • Diffusion-based lifting becomes a practical option for occluded settings, since the model can sample multiple plausible 3D poses rather than committing to one.
  • Hypergraph convolution can serve as a general denoiser for structured outputs, since many body and object reconstruction tasks have multi-joint correlations that pairwise graphs under-represent.
  • The adjustable number of diffusion steps gives a simple accuracy-versus-compute dial, useful for deployment on devices with different computational budgets.

Reading between the lines

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

  • My editorial inference: the headline result will only be convincing if the benchmark comparison holds the 2D detector fixed; otherwise detector quality, not the architecture, could explain the gains.
  • A direct testable extension is to swap the multi-granularity HyperGCN denoiser for a standard graph-convolution denoiser of equal parameter count; if the gap is small, the hyperedges are not the active ingredient.
  • Because diffusion samplers can be distilled, HyperDiff's compute-accuracy tradeoff might be pushed further by compressing the multi-step denoising process into a single-step denoiser, making the method more deployable.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 3 minor

Summary. The manuscript proposes HyperDiff, a diffusion-based 2D-to-3D human pose estimator whose denoiser is a multi-granularity HyperGCN. The abstract claims state-of-the-art performance on Human3.6M and MPI-INF-3DHP, and a flexible compute/accuracy tradeoff. The supplied full text is almost entirely corrupted: the readable abstract contains no numerical results, protocol descriptions, or ablation summaries, and the remainder of the file is unreadable replacement characters, including a stray arXiv header for a physics paper. Thus the experimental support for the main claims cannot be inspected.

Significance. If the claims are accurate and the ablations isolate the proposed components, HyperDiff would be a reasonable contribution: it combines diffusion-based uncertainty modeling for depth/occlusion with a multi-granularity hypergraph denoiser for joint correlations, and an inference-time compute/accuracy tradeoff is practically useful. However, because no quantitative results, evaluation protocol, detector, or ablations are readable, I cannot confirm the contribution's significance.

major comments (3)
  1. [Abstract; Experiments] The abstract's central sentence—'achieves state-of-the-art performance'—is unsubstantiated. The paper gives no MPJPE/PA-MPJPE numbers, no protocol (P1 vs P2, joint count), no 2D detector, no alignment/refinement settings, and no baseline table. In 2D-to-3D lifting, final error is largely controlled by 2D detector quality; without these details the SOTA claim is not comparable or verifiable.
  2. [Full text (unreadable)] The supplied full text consists of replacement characters and contains a header from arXiv:2508.14430v1 [physics.atom-ph]. The methods, equations, tables, and ablations are all unrecoverable. As presented, the paper does not contain the evidence needed to evaluate the proposed HyperGCN construction, diffusion schedule, loss function, or sampling procedure, nor the claimed advantages for depth ambiguity and occlusion.
  3. [Ablations (unreadable)] No ablation separates the multi-granularity hypergraph from plain HyperGCN, diffusion sampling steps, or model capacity. Therefore the mechanistic claim that multi-granularity hyperedges cause the reported improvement is not supported. Even if a table were recoverable, the protocol/detector comparability issue would need to be resolved.
minor comments (3)
  1. [Full text] The text needs to be regenerated/de-encoded; the current file is unusable for review.
  2. [Header] The arXiv physics header (2508.14430v1) should be removed; it appears to belong to a different paper.
  3. [Abstract] The abstract should include at least one concrete number and the evaluation protocol (detector, protocol, metric) to support the SOTA claim.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity identified: the central SOTA claim is evaluated on external benchmarks and no prediction reduces to a fitted input or self-citation chain.

full rationale

The paper's headline claim is that HyperDiff 'achieves state-of-the-art performance on the Human3.6M and MPI-INF-3DHP datasets.' These are standard external benchmarks whose labels and protocols are not authored by the present paper, so the central result is not circular by construction. The supplied full text is heavily corrupted, so the derivation equations, ablations, and comparison settings cannot be inspected; however, an evidentiary gap or an unstated evaluation protocol is a correctness/verifiability concern, not circularity. There is no visible step where a parameter is fitted to a subset and then a closely related quantity is called a prediction, no visible self-citation used as the load-bearing justification for a design choice, no imported uniqueness theorem, and no visible renaming of a known result as a new contribution. Because the hard rule requires quoting an actual reduction or dependence, and none can be exhibited from the available text, the appropriate finding is no significant circularity.

Assumptions & free parameters 3 free parameters · 4 assumptions · 1 invented entities

The abstract-level view is that the paper rests on standard ML assumptions (diffusion training, benchmark validity, upstream 2D detection) plus its own architectural postulate (multi-granularity hypergraph denoising). No numeric values are visible to count fitted parameters beyond typical hyperparameter choices.

free parameters (3)
  • number of diffusion inference (sampling) steps
    Typical hand-tuned hyperparameter governing the performance/efficiency tradeoff the abstract promises; value not stated.
  • granularity scales of hypergraph construction
    Number and grouping of multi-granularity hyperedges chosen by design; not specified in the abstract.
  • loss weights and training hyperparameters
    Diffusion denoising loss versus pose loss weighting and learning schedule; unspecified in the abstract.
assumptions (4)
  • domain assumption 2D keypoints are available as input (2D-to-3D lifting setting)
    The abstract frames the task as a '2D-to-3D lifting process'; this requires a 2D detector upstream, and the quality of that detector is not stated.
  • domain assumption Human3.6M and MPI-INF-3DHP are valid external benchmarks with accepted protocols
    The SOTA claim rests on these datasets being standard and fairly compared benchmarks; no protocol details appear in the abstract.
  • domain assumption The denoising diffusion objective is a valid training target for the 3D pose posterior
    The paper assumes diffusion training captures pose uncertainty in a way that helps depth disambiguation; this is standard in prior diffusion HPE works.
  • ad hoc to paper Multi-granularity hypergraph hyperedges provide the right inductive bias for joint correlations
    Core architectural postulate of the paper; gains attributed to it would need ablations to support, which are unreadable here.
invented entities (1)
  • Multi-granularity hypergraph denoiser (HyperGCN)
    purpose: Serve as the diffusion denoiser, modeling high-order correlations across joint groupings at several scales
    New architectural component introduced by the paper; its benefit is claimed via benchmark results that are not visible in this review.

how reviews work

0 comments
Cite this review

Pith. "Pith review of HyperDiff: Hypergraph Guided Diffusion Model for 3D Human Pose Estimation." pith.science (2026). https://pith.science/paper/U7ROJ6XB

@misc{pith2026250814431,
  author       = {Pith},
  title        = {Pith review of: HyperDiff: Hypergraph Guided Diffusion Model for 3D Human Pose Estimation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/U7ROJ6XB}},
  note         = {Machine review of arXiv:2508.14431}
}
read the original abstract

Monocular 3D human pose estimation (HPE) often encounters challenges such as depth ambiguity and occlusion during the 2D-to-3D lifting process. Additionally, traditional methods may overlook multi-scale skeleton features when utilizing skeleton structure information, which can negatively impact the accuracy of pose estimation. To address these challenges, this paper introduces a novel 3D pose estimation method, HyperDiff, which integrates diffusion models with HyperGCN. The diffusion model effectively captures data uncertainty, alleviating depth ambiguity and occlusion. Meanwhile, HyperGCN, serving as a denoiser, employs multi-granularity structures to accurately model high-order correlations between joints. This improves the model's denoising capability especially for complex poses. Experimental results demonstrate that HyperDiff achieves state-of-the-art performance on the Human3.6M and MPI-INF-3DHP datasets and can flexibly adapt to varying computational resources to balance performance and efficiency.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

46 extracted references · 42 canonical work pages

  1. [1]

    Y. Chen, Z. Tu, L. Ge, D. Zhang, R. Chen, and J. Yuan, ``So-handnet: Self-organizing network for 3d hand pose estimation with semi-supervised learning,'' in Proceedings of the IEEE/CVF international conference on computer vision , pp. 6961--6970, 2019

    work page 2019

  2. [2]

    Y. Chen, Z. Tu, D. Kang, L. Bao, Y. Zhang, X. Zhe, R. Chen, and J. Yuan, ``Model-based 3d hand reconstruction via self-supervised learning,'' in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pp. 10451--10460, 2021

    work page 2021

  3. [3]

    Y. Chen, Z. Tu, D. Kang, R. Chen, L. Bao, Z. Zhang, and J. Yuan, ``Joint hand-object 3d reconstruction from a single image with cross-branch feature fusion,'' IEEE Transactions on Image Processing , vol. 30, pp. 4008--4021, 2021

    work page 2021

  4. [4]

    J. Wang, H. Xu, J. Xu, S. Liu, and X. Wang, ``Synthesizing long-term 3d human motion and interaction in 3d scenes,'' in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pp. 9401--9411, 2021

    work page 2021

  5. [5]

    Hassan, P

    M. Hassan, P. Ghosh, J. Tesch, D. Tzionas, and M. J. Black, ``Populating 3d scenes by learning human-scene interaction,'' in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pp. 14708--14718, 2021

    work page 2021

  6. [6]

    E. Ng, D. Xiang, H. Joo, and K. Grauman, ``You2me: Inferring body pose in egocentric video via first and second person interactions,'' in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pp. 9890--9900, 2020

    work page 2020

  7. [7]

    Q. Lu, W. Han, J. Ling, M. Wang, H. Chen, B. Varadarajan, and P. Covington, ``Kemp: Keyframe-based hierarchical end-to-end deep model for long-term trajectory prediction,'' in 2022 International Conference on Robotics and Automation (ICRA) , pp. 646--652, IEEE, 2022

    work page 2022

  8. [8]

    Newell, K

    A. Newell, K. Yang, and J. Deng, ``Stacked hourglass networks for human pose estimation,'' in Computer Vision--ECCV 2016: 14th European Conference, Amsterdam, The Netherlands, October 11-14, 2016, Proceedings, Part VIII 14 , pp. 483--499, Springer, 2016

    work page 2016

Show all 46 references
  1. [9]

    S.-E. Wei, V. Ramakrishna, T. Kanade, and Y. Sheikh, ``Convolutional pose machines,'' in Proceedings of the IEEE conference on Computer Vision and Pattern Recognition , pp. 4724--4732, 2016

    2016

  2. [10]

    Y. Chen, Z. Wang, Y. Peng, Z. Zhang, G. Yu, and J. Sun, ``Cascaded pyramid network for multi-person pose estimation,'' in Proceedings of the IEEE conference on computer vision and pattern recognition , pp. 7103--7112, 2018

    2018

  3. [11]

    K. Sun, B. Xiao, D. Liu, and J. Wang, ``Deep high-resolution representation learning for human pose estimation,'' in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pp. 5693--5703, 2019

    2019

  4. [12]

    J. Ho, A. Jain, and P. Abbeel, ``Denoising diffusion probabilistic models,'' Advances in neural information processing systems , vol. 33, pp. 6840--6851, 2020

    2020

  5. [13]

    J. Song, C. Meng, and S. Ermon, ``Denoising diffusion implicit models,'' arXiv preprint arXiv:2010.02502 , 2020

    2010 arXiv

  6. [14]

    W. Shan, Z. Liu, X. Zhang, Z. Wang, K. Han, S. Wang, S. Ma, and W. Gao, ``Diffusion-based 3d human pose estimation with multi-hypothesis aggregation,'' in Proceedings of the IEEE/CVF International Conference on Computer Vision , pp. 14761--14771, 2023

    2023

  7. [15]

    J. Xu, Y. Guo, and Y. Peng, ``Finepose: Fine-grained prompt-driven 3d human pose estimation via diffusion models,'' in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pp. 561--570, 2024

    2024

  8. [16]

    Q. Cai, X. Hu, S. Hou, L. Yao, and Y. Huang, ``Disentangled diffusion-based 3d human pose estimation with hierarchical spatial and temporal denoiser,'' in Proceedings of the AAAI Conference on Artificial Intelligence , vol. 38, pp. 882--890, 2024

    2024

  9. [17]

    W. Li, M. Liu, H. Liu, T. Guo, T. Wang, H. Tang, and N. Sebe, ``Graphmlp: A graph mlp-like architecture for 3d human pose estimation,'' Pattern Recognition , vol. 158, p. 110925, 2025

    2025

  10. [18]

    L. Shi, Y. Zhang, J. Cheng, and H. 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 , pp. 12026--12035, 2019

    2019

  11. [19]

    S. Yan, Y. Xiong, and D. Lin, ``Spatial temporal graph convolutional networks for skeleton-based action recognition,'' in Proceedings of the AAAI conference on artificial intelligence , vol. 32, 2018

    2018

  12. [20]

    Bruna, W

    J. Bruna, W. Zaremba, A. Szlam, and Y. LeCun, ``Spectral networks and locally connected networks on graphs,'' arXiv preprint arXiv:1312.6203 , 2013

    2013 arXiv

  13. [21]

    T. N. Kipf and M. Welling, ``Semi-supervised classification with graph convolutional networks,'' arXiv preprint arXiv:1609.02907 , 2016

    2016 arXiv

  14. [22]

    W. Hu, C. Zhang, F. Zhan, L. Zhang, and T.-T. Wong, ``Conditional directed graph convolution for 3d human pose estimation,'' in Proceedings of the 29th ACM International Conference on Multimedia , pp. 602--611, 2021

    2021

  15. [23]

    B. X. Yu, Z. Zhang, Y. Liu, S.-h. Zhong, Y. Liu, and C. W. Chen, ``Gla-gcn: Global-local adaptive graph convolutional network for 3d human pose estimation from monocular video,'' in Proceedings of the IEEE/CVF International Conference on Computer Vision , pp. 8818--8829, 2023

    2023

  16. [24]

    W. Zhao, W. Wang, and Y. Tian, ``Graformer: Graph-oriented transformer for 3d pose estimation,'' in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pp. 20438--20447, 2022

    2022

  17. [25]

    Fey and J

    M. Fey and J. E. Lenssen, ``Fast graph representation learning with PyTorch Geometric ,'' in ICLR Workshop on Representation Learning on Graphs and Manifolds , 2019

    2019

  18. [26]

    Yadati, M

    N. Yadati, M. Nimishakavi, P. Yadav, V. Nitin, A. Louis, and P. Talukdar, ``Hypergcn: A new method for training graph convolutional networks on hypergraphs,'' Advances in neural information processing systems , vol. 32, 2019

    2019

  19. [27]

    J. Wei, Y. Wang, M. Guo, P. Lv, X. Yang, and M. Xu, ``Dynamic hypergraph convolutional networks for skeleton-based action recognition,'' 2021

    2021

  20. [28]

    Y. Zhu, G. Huang, X. Xu, Y. Ji, and F. Shen, ``Selective hypergraph convolutional networks for skeleton-based action recognition,'' in Proceedings of the 2022 international conference on multimedia retrieval , pp. 518--526, 2022

    2022

  21. [29]

    Y. Feng, H. You, Z. Zhang, R. Ji, and Y. Gao, ``Hypergraph neural networks,'' in Proceedings of the AAAI conference on artificial intelligence , vol. 33, pp. 3558--3565, 2019

    2019

  22. [30]

    Zou and W

    Z. Zou and W. Tang, ``Modulated graph convolutional network for 3d human pose estimation,'' in Proceedings of the IEEE/CVF international conference on computer vision , pp. 11477--11487, 2021

    2021

  23. [31]

    S. Bai, F. Zhang, and P. H. Torr, ``Hypergraph convolution and hypergraph attention,'' Pattern Recognition , vol. 110, p. 107637, 2021

    2021

  24. [32]

    H. Li, B. Shi, W. Dai, H. Zheng, B. Wang, Y. Sun, M. Guo, C. Li, J. Zou, and H. Xiong, ``Pose-oriented transformer with uncertainty-guided refinement for 2d-to-3d human pose estimation,'' in Proceedings of the AAAI conference on artificial intelligence , vol. 37, pp. 1296--1304, 2023

    2023

  25. [33]

    J. Gong, L. G. Foo, Z. Fan, Q. Ke, H. Rahmani, and J. Liu, ``Diffpose: Toward more reliable 3d pose estimation,'' in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pp. 13041--13051, 2023

    2023

  26. [34]

    M. T. Hassan and A. B. Hamza, ``Regular splitting graph network for 3d human pose estimation,'' IEEE Transactions on Image Processing , vol. 32, pp. 4212--4222, 2023

    2023

  27. [35]

    W. Wang, J. Xiao, C. Wang, W. Liu, Z. Wang, and L. Chen, `` Di ^ 2 Pose : Discrete diffusion model for occluded 3d human pose estimation,'' Advances in Neural Information Processing Systems , vol. 37, pp. 98717--98741, 2024

    2024

  28. [36]

    F. Zhou, J. Yin, and P. Li, ``Lifting by image--leveraging image cues for accurate 3d human pose estimation,'' in Proceedings of the AAAI Conference on Artificial Intelligence , vol. 38, pp. 7632--7640, 2024

    2024

  29. [37]

    Xu and W

    T. Xu and W. Takano, ``Graph stacked hourglass networks for 3d human pose estimation,'' in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pp. 16105--16114, 2021

    2021

  30. [38]

    H. Li, B. Shi, W. Dai, Y. Chen, B. Wang, Y. Sun, M. Guo, C. Li, J. Zou, and H. Xiong, ``Hierarchical graph networks for 3d human pose estimation,'' arXiv preprint arXiv:2111.11927 , 2021

    2021 arXiv

  31. [39]

    Zhang, C

    S. Zhang, C. Wang, L. Nie, H. Yao, Q. Huang, and Q. Tian, ``Learning enriched hop-aware correlation for robust 3d human pose estimation,'' International Journal of Computer Vision , vol. 131, no. 6, pp. 1566--1583, 2023

    2023

  32. [40]

    Zheng, S

    C. Zheng, S. Zhu, M. Mendieta, T. Yang, C. Chen, and Z. Ding, ``3d human pose estimation with spatial and temporal transformers,'' in Proceedings of the IEEE/CVF international conference on computer vision , pp. 11656--11665, 2021

    2021

  33. [41]

    W. Li, H. Liu, H. Tang, P. Wang, and L. Van Gool, ``Mhformer: Multi-hypothesis transformer for 3d human pose estimation,'' in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pp. 13147--13156, 2022

    2022

  34. [42]

    Ionescu, D

    C. Ionescu, D. Papava, V. Olaru, and C. Sminchisescu, ``Human3. 6m: Large scale datasets and predictive methods for 3d human sensing in natural environments,'' IEEE transactions on pattern analysis and machine intelligence , vol. 36, no. 7, pp. 1325--1339, 2013

    2013

  35. [43]

    Z. Tang, Z. Qiu, Y. Hao, R. Hong, and T. Yao, ``3d human pose estimation with spatio-temporal criss-cross attention,'' in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pp. 4790--4799, 2023

    2023

  36. [44]

    J. Peng, Y. Zhou, and P. Mok, ``Ktpformer: Kinematics and trajectory prior knowledge-enhanced transformer for 3d human pose estimation,'' in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pp. 1123--1132, 2024

    2024

  37. [45]

    Martinez, R

    J. Martinez, R. Hossain, J. Romero, and J. J. Little, ``A simple yet effective baseline for 3d human pose estimation,'' in Proceedings of the IEEE international conference on computer vision , pp. 2640--2649, 2017

    2017

  38. [46]

    S. Li, L. Ke, K. Pratama, Y.-W. Tai, C.-K. Tang, and K.-T. Cheng, ``Cascaded deep monocular 3d human pose estimation with evolutionary training data,'' in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pp. 6173--6183, 2020

    2020

Pith tools

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