Pith. sign in

REVIEW 5 major objections 6 minor 52 references

Optimizing Local-Global Dependencies for Accurate 3D Human Pose Estimation

T0 review · 5 major / 6 minor · reviewed 2026-08-10 · deepseek-v4-flash

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

Pith's one-line read The paper claims that a dual-stream model combining skeleton selective refine attention with a global transformer achieves state-of-the-art 3D pose accuracy, with a 12.8 mm MPJPE on Human3.6M ground-truth 2D poses.

desk verdict Reasonable incremental architecture with a strong SOTA claim, but the missing evaluation protocol makes the headline numbers unverifiable as written. read the letter →

arxiv 2412.19676 v1 pith:HQY7LISG submitted 2024-12-27 cs.CV

Guangsheng Xu , Guoyi Zhang , Lejia Ye , Shuwei Gan , Xiaohu Zhang , Xia Yang This is my paper · ORCID
classification cs.CV
keywords 3Dhumanposeestimationskeletonselectiverefineattentionlarge-kernelhybridtransformerdual-streamnetworkmotionrepresentationmeshrecoveryMPJPE
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 tries to show that 3D human pose estimation from 2D skeletons improves when a model combines two parallel feature streams: a transformer that captures global spatio-temporal dependencies and a lightweight module, SSRFormer, that captures fine-grained local joint and motion details with a skeleton selective refine attention mechanism. If correct, the model sets a new accuracy standard on the two main benchmarks, with 37.4 mm mean per-joint position error on Human3.6M using detected 2D poses, 12.8 mm using ground-truth 2D poses at 243 input frames, and 13.2 mm on MPI-INF-3DHP. The paper also claims the learned motion representation transfers to human mesh recovery, improving on MotionBERT in vertex and joint error. The reason to care is that this would show that the bottleneck in 2D-to-3D lifting is not only long-range context but also local structure, and that a carefully designed large-kernel attention can supply that local structure at low parameter cost.

What carries the argument

The skeleton selective refine attention (SSRA) mechanism, an irregular large-kernel attention built from depth-wise and dilated depth-wise convolutions decomposed into separable 1D kernels. By reshaping the input so the long kernel side aligns with either the joint dimension (spatial SSRFormer) or the time dimension (temporal SSRFormer), SSRA assigns different weights to spatial and temporal local aggregation. A pointwise convolution produces an attention map that modulates the input through a Hadamard product, making the local stream input-adaptive rather than a fixed convolution. This local stream runs in parallel with the global transformer stream, and adaptive fusion weights combine the two.

What would settle it

Run a single, explicitly stated evaluation on Human3.6M, for example central-frame MPJPE with a sliding window and fixed stride, for SSR-STF and the top baselines MotionAGFormer, MotionBERT, and KTPFormer, and check whether the reported margins, particularly the 4.5 mm ground-truth-2D gain, persist under identical frame scoring and sampling.

Watch

Extended reading notes

Core claim

SSR-STF is a dual-stream network in which each block runs an SSRFormer and an STFormer in parallel. The STFormer stream uses spatio-temporal criss-cross attention to model global dependencies, while the SSRFormer stream applies skeleton selective refine attention to selectively refine local dependencies along either the joint or time dimension, and the two streams are combined by an adaptive, learned weighted fusion. On Human3.6M the paper reports P1 errors of 37.4 mm with detected 2D poses and 12.8 mm with ground-truth 2D poses at T=243, and on MPI-INF-3DHP an MPJPE of 13.2 mm with PCK of 99.2 percent. In the ground-truth-2D Human3.6M setting, it claims a 4.5 mm improvement over MotionAGFormer, the previous best, and the same motion representation, when regressed to SMPL parameters, reportedly lowers mesh recovery error from 75.7 to 68.4 mm MPVE relative to MotionBERT.

Load-bearing premise

The reported state-of-the-art result depends on the evaluation protocol for SSR-STF being identical to the protocols used for the cited baselines; the paper never states whether MPJPE is averaged over every frame in a clip or only the central frame, nor how input windows are sampled.

Editorial extensions

If this is right

  • If the central claim holds, the practical upper bound of 2D-to-3D lifting improves substantially: 12.8 mm MPJPE from ground-truth 2D poses at T=243, about 26 percent below the previous best reported in the paper.
  • Hybrid local-global architectures would be validated over pure global transformers, giving a design pattern of parallel streams with adaptive fusion rather than sequential stacking.
  • The learned motion representation transfers to SMPL-based mesh recovery, suggesting the same backbone can support downstream human-centric tasks.
  • On MPI-INF-3DHP, the model generalizes across capture settings with 13.2 mm MPJPE and 99.2 percent PCK using ground-truth 2D poses, indicating robustness to background and camera variation.

Reading between the lines

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

  • A natural testable extension is to apply SSRA to other skeleton-based tasks, such as action recognition, where local joint correlations and temporal motion details matter as much as long-range context.
  • Because the paper does not state whether MPJPE is scored on the central frame only or on all frames in the clip, a matched-protocol re-evaluation could narrow the reported margin over baselines; this is an inference about the comparison, not a claim the paper makes.
  • The irregular kernel design suggests an architecture that adapts its receptive field to sequence length or joint count; one could vary the kernel shape per block and measure the accuracy-efficiency frontier.
  • The large gains on ground-truth 2D poses indicate the lifting model itself, not the 2D detector, drives the improvement; testing the same backbone with newer 2D detectors could show whether the margin compounds.
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

5 major / 6 minor

Summary. The paper proposes SSR-STF, a dual-stream 2D-to-3D human pose estimation model that combines a Transformer stream (STFormer) capturing global spatio-temporal dependencies with a newly proposed SSRFormer stream using skeleton selective refine attention (SSRA), an irregular large-kernel attention mechanism, for local dependencies. The two streams are adaptively fused, and the model is trained to predict all T frames of an input window under a position plus velocity loss. Experiments on Human3.6M and MPI-INF-3DHP report state-of-the-art MPJPE/P-MPJPE results, including 37.4 mm on Human3.6M with detected 2D poses, 12.8 mm with ground-truth 2D poses at T=243, and 13.2 mm on MPI-INF-3DHP, together with ablations and a downstream human mesh recovery experiment showing improvements over MotionBERT.

Significance. If the reported numbers are obtained under a protocol comparable to that of the cited baselines, the work makes a solid empirical contribution: it presents a simple, modular way to complement global Transformer attention with local large-kernel attention in skeleton sequences, it includes extensive comparisons across two benchmarks and several temporal window sizes, it releases code, and it evaluates the learned motion representation on mesh recovery. The architecture is internally consistent and the ablation study supports the value of the dual-stream adaptive-fusion design. The main significance hinges on the comparability of the evaluation protocol, which is not currently documented.

major comments (5)
  1. [Section IV-A/IV-B, Tables I-III] The central claim of state-of-the-art accuracy is not verifiable because the evaluation protocol for the temporal window is never specified. Equation (12) trains the model to predict all T frames, so SSR-STF naturally produces T poses per input clip, whereas some baselines (PoseFormer, MHFormer, P-STMO) are central-frame predictors and others (MixSTE, MotionBERT) are sequence-to-sequence. The paper does not state whether reported MPJPE is computed on the central frame only, on all frames, with a stride-1 sliding window, or with non-overlapping windows. All-frame evaluation with overlapping windows can average multiple predictions per frame and is not directly comparable to center-frame evaluation. Please specify the exact protocol for every table and, ideally, re-evaluate the compared baselines under the identical protocol.
  2. [Table IV] The parameter counts in Table IV are internally inconsistent: the configurations with Ch=256 and Ch=512 both report 36.7M parameters, yet Ch is the motion embedding dimension used by the final linear layer and the regression head, so changing it should change the parameter count. The Ch=1024 row also increases by only 0.2M, which is implausible if the projection and head scale with Ch. Please correct the reported counts or explain why these parameters are shared or omitted.
  3. [Table III and Section IV-D] There is a citation/reference error in Table III: the MotionAGFormer row is cited as [28], but [28] in the reference list is KTPFormer, while MotionAGFormer is [8]. The text in Section IV-D also lists comparison baselines as [8], [17], [51], which does not match the table. This must be corrected, and the per-action breakdown for MotionAGFormer in Table II should be provided or explicitly marked as unavailable.
  4. [Tables I-VI] No variance or repeated-run information is given for any reported number, yet several state-of-the-art claims rest on differences of about 1 mm or less (e.g., 37.4 vs. 38.4 for MotionAGFormer in Table I, and 31.7 vs. 31.9 in P-MPJPE). Without error bars, seeds, or at least a statement about the number of runs, the reader cannot assess whether these gaps are meaningful. Please add variance information or, if only one run was performed, state this explicitly and discuss the sensitivity of the conclusions.
  5. [Section IV-E, Table VI] The text discussing kernel shapes says that on Human3.6M the 35x11 kernel gives a 0.3 mm lower P1 error than the 11x1 kernel, but Table VI is titled as reporting results on MPI-INF-3DHP only and does not contain Human3.6M values. Please clarify which dataset Table VI reports, and either add the Human3.6M kernel comparisons or remove the unsupported claim.
minor comments (6)
  1. [Table I (bottom)] In the P-MPJPE block, the MotionBERT row lists 22.0 for WalkD. and 31.7 for Walk, which appears to be a swapped pair compared with the values reported in the original MotionBERT paper; please verify these entries against the source.
  2. [Fig. 3] Figure 3 contains the Chinese text '取名为不规则LK' and the undefined label 'LKFormer'; this is inconsistent with the English exposition and should be cleaned up before publication.
  3. [Abstract and Section I] The code link is given as a URL in the abstract but appears as the placeholder 'SSR-STF' in the full text; please use the actual repository URL consistently.
  4. [Section III-B3, Eq. (4)] The order of operators in Eq. (4) (DWD2(DWD1(DW2(DW1(X))))) is not clearly aligned with the surrounding text describing DW1/DW2 and DWD1/DWD2, and the kernel sizes in the text and in Fig. 3 do not match notationally; please harmonize the notation and the order of operations.
  5. [Reference list] Reference [5] is incomplete ('A. Vaswani, Attention is all you need', without venue or year), reference [35] lists the author as 'D. Alexey', and several other entries have inconsistent formatting; please normalize the bibliography.
  6. [Table II] The MotionAGFormer row in Table II reports only the average (17.3 mm) with all per-action entries missing; either provide the full row or state that the per-action numbers are not available from the source.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: SSR-STF's SOTA claims are empirical, benchmarked against external datasets with supervised losses, and no fitted quantity is renamed as a prediction.

full rationale

The paper's derivation chain is an architecture plus supervised training on external benchmarks. The claimed outputs (MPJPE 37.4 mm on Human3.6M with detected 2D poses, 12.8 mm with GT 2D poses, and 13.2 mm on MPI-INF-3DHP) are measured on held-out subjects against ground-truth 3D poses, not derived from the model's own inputs or hyperparameters. Eq. (12) trains the model toward external ground truth X and velocity difference ΔX, and the adaptive fusion weights in Eq. (11) are internal learned parameters, not predicted quantities that are later reported as results. The mesh-recovery experiment (Table VII) regresses SMPL parameters from the learned motion representation E and is scored against SMPL ground truth, so the target is not defined by the model itself. The only overlapping-author citation, reference [43], supports a general related-work claim about large-kernel convolutional networks competing with Transformers; it is not load-bearing for any of the paper's accuracy claims. The evaluation-protocol ambiguity noted in the reader's take is a comparability and correctness risk, not a circularity, because it does not make any reported number equal to an input by construction. No self-definitional, fitted-input-called-prediction, self-citation-load-bearing, uniqueness-imported, ansatz-smuggled, or renaming-known-result pattern is present.

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

The paper contributes an architecture and empirical results; it relies on standard benchmark assumptions and on the equivalence of evaluation protocols. The main free parameters are architecture hyperparameters chosen by validation; the velocity-loss weight is not reported. No new entities are introduced.

free parameters (5)
  • Number of blocks N = 12
    Selected from {6,8,10,12,14,16} by validation on Human3.6M (Table IV).
  • Channel width C = 256
    Selected from {128,256} by validation; C=256 with N=12 gives best MPJPE.
  • Motion embedding dimension Ch = 512
    Selected from {256,512,1024}; 512 gives the best MPJPE in Table IV.
  • SSRA kernel shape = 35x11 (Human3.6M), 11x1 (MPI-INF-3DHP)
    Chosen per dataset from Table VI; different kernel shapes give different errors.
  • Velocity loss weight lambda_deltaX = not reported
    Introduced in Eq.13 but its value is never given; it affects the smoothness/accuracy tradeoff.
assumptions (5)
  • standard math A k x k convolution can be decomposed into a (2d-1)x(2d-1) depthwise conv, a depthwise dilated conv, and a 1x1 conv (VAN), with further 1D separation (LSKA).
    Used in Eq.4 and Figure 3 to build the SSRA module; not re-derived in this paper.
  • domain assumption Ground-truth 2D poses on Human3.6M and MPI-INF-3DHP represent the upper bound of 2D-to-3D lifting models.
    Used to position the results in Tables II and III.
  • domain assumption The evaluation protocol for all compared methods is identical, including frame sampling, stride, and whether central or all frames are scored.
    The paper does not describe the protocol, so the state-of-the-art claim rests on this unstated premise.
  • domain assumption Stacked Hourglass 2D detections on Human3.6M are the same ones used by the compared baselines.
    The paper says 'in accordance with [18]' but does not verify that all baselines used the same detections.
  • domain assumption MPI-INF-3DHP evaluation uses ground-truth 2D poses as in the cited baselines [8], [17], [51].
    Stated in Section IV-B; necessary for a fair comparison.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Optimizing Local-Global Dependencies for Accurate 3D Human Pose Estimation." pith.science (2026). https://pith.science/paper/HQY7LISG

@misc{pith2026241219676,
  author       = {Pith},
  title        = {Pith review of: Optimizing Local-Global Dependencies for Accurate 3D Human Pose Estimation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/HQY7LISG}},
  note         = {Machine review of arXiv:2412.19676}
}
read the original abstract

Transformer-based methods have recently achieved significant success in 3D human pose estimation, owing to their strong ability to model long-range dependencies. However, relying solely on the global attention mechanism is insufficient for capturing the fine-grained local details, which are crucial for accurate pose estimation. To address this, we propose SSR-STF, a dual-stream model that effectively integrates local features with global dependencies to enhance 3D human pose estimation. Specifically, we introduce SSRFormer, a simple yet effective module that employs the skeleton selective refine attention (SSRA) mechanism to capture fine-grained local dependencies in human pose sequences, complementing the global dependencies modeled by the Transformer. By adaptively fusing these two feature streams, SSR-STF can better learn the underlying structure of human poses, overcoming the limitations of traditional methods in local feature extraction. Extensive experiments on the Human3.6M and MPI-INF-3DHP datasets demonstrate that SSR-STF achieves state-of-the-art performance, with P1 errors of 37.4 mm and 13.2 mm respectively, outperforming existing methods in both accuracy and generalization. Furthermore, the motion representations learned by our model prove effective in downstream tasks such as human mesh recovery. Codes are available at https://github.com/poker-xu/SSR-STF.

Figures

Figures reproduced from arXiv: 2412.19676 by the authors.

Figure 1
Figure 1. MPJPE Comparison on Human3.6M dataset [11] (lower is better). The horizontal and vertical axes represent MPJPE using detected 2D poses and GT 2D poses as inputs, respectively. Using GT 2D poses as input can evaluate the performance upper bound of 2D-to-3D lifting models. The number of parameters for each algorithm is also provided. Our model strikes a balance between performance and parameter efficiency, achieving a… view at source ↗
Figure 2
Figure 2. (a) The overall architecture of SSR-STF, which is characterized by N dual-stream spatio-temporal blocks, includes one stream leveraging SSRFormers and the other employing STFormers. (b) Network structure of the Spatial/Temporal SSRFormer. SSRFormer employs skeleton selective refine attention mechanism (i.e., SSRA) to capture the local spatio-temporal features of skeleton sequences. (c) Network structure of the STFor… view at source ↗
Figure 3
Figure 3. SSRFormer. We employ skeleton selective refine attention mechanism to extract the spatio-temporal local features of 2D joints, as illustrated by the SSRFormer with a kernel size of k1 × k2. in either the temporal or spatial dimension, thereby assigning different weights to the feature aggregation relationships in each dimension. The heart of SSR is the SSRA, as shown in Fig.3, which selectively captures local tempor… view at source ↗
Figures from the paper (2 more)
Figure 4
Figure 4. Figure 4: The MPJPE distribution on Human3.6M testset, with the estimated 2D [PITH_FULL_IMAGE:figures/full_fig_p006_4.png]
Figure 5
Figure 5. Figure 5: Qualitative comparisons of 3D pose estimation by MixSTE [16], [PITH_FULL_IMAGE:figures/full_fig_p008_5.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

52 extracted references · 33 canonical work pages

  1. [28]

    Ktpformer: Kinematics and trajectory prior knowledge-enhanced transformer for 3d human pose estimation,

    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, 2024, pp. 1123–1132

    work page 2024

  2. [8]

    Motionagformer: Enhancing 3d human pose estimation with a transformer-gcnformer network,

    S. Mehraban, V . Adeli, and B. Taati, “Motionagformer: Enhancing 3d human pose estimation with a transformer-gcnformer network,” in Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision, 2024, pp. 6920–6930

    work page 2024

  3. [17]

    3d human pose estimation with spatio-temporal criss-cross attention,

    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, 2023, pp. 4790–4799

    work page 2023

  4. [51]

    PoseFormerV2: Exploring frequency domain for efficient and robust 3d human pose estimation,

    Q. Zhao, C. Zheng, M. Liu, P. Wang, and C. Chen, “PoseFormerV2: Exploring frequency domain for efficient and robust 3d human pose estimation,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) , June 2023, pp. 8877–8886

    work page 2023

  5. [1]

    Pose estimation and adaptive robot behaviour for human-robot interaction,

    M. Svenstrup, S. Tranberg, H. J. Andersen, and T. Bak, “Pose estimation and adaptive robot behaviour for human-robot interaction,” in2009 IEEE International Conference on Robotics and Automation . IEEE, 2009, pp. 3571–3576

    work page 2009

  6. [2]

    The progress of human pose estimation: A survey and taxonomy of models applied in 2d human pose estimation,

    T. L. Munea, Y . Z. Jembre, H. T. Weldegebriel, L. Chen, C. Huang, and C. Yang, “The progress of human pose estimation: A survey and taxonomy of models applied in 2d human pose estimation,” IEEE Access, vol. 8, pp. 133 330–133 348, 2020

    work page 2020

  7. [3]

    Vnect: Real-time 3d human pose estimation with a single rgb camera,

    D. Mehta, S. Sridhar, O. Sotnychenko, H. Rhodin, M. Shafiei, H.-P. Seidel, W. Xu, D. Casas, and C. Theobalt, “Vnect: Real-time 3d human pose estimation with a single rgb camera,” Acm transactions on graphics (tog), vol. 36, no. 4, pp. 1–14, 2017

    work page 2017

  8. [4]

    Weakly supervised multi-modal 3d human body pose estimation for autonomous driving,

    P. Bauer, A. Bouazizi, U. Kressel, and F. B. Flohr, “Weakly supervised multi-modal 3d human body pose estimation for autonomous driving,” in 2023 IEEE Intelligent Vehicles Symposium (IV) . IEEE, 2023, pp. 1–7

    work page 2023

Show all 52 references
  1. [5]

    Attention is all you need,

    A. Vaswani, “Attention is all you need,” Advances in Neural Information Processing Systems, 2017

    2017

  2. [6]

    Graformer: Graph-oriented transformer for 3d pose estimation,

    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 , 2022, pp. 20 438–20 447

    2022

  3. [7]

    Pose-oriented transformer with uncertainty- guided refinement for 2d-to-3d human pose estimation,

    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, no. 1, 2023, ...

    2023

  4. [9]

    Measuring and relieving the over-smoothing problem for graph neural networks from the topological view,

    D. Chen, Y . Lin, W. Li, P. Li, J. Zhou, and X. Sun, “Measuring and relieving the over-smoothing problem for graph neural networks from the topological view,” inProceedings of the AAAI conference on artificial intelligence, vol. 34, no. 04, 2020, pp. 3438–3445

    2020

  5. [10]

    Skeleton-based human ac- tion recognition via large-kernel attention graph convolutional network,

    Y . Liu, H. Zhang, Y . Li, K. He, and D. Xu, “Skeleton-based human ac- tion recognition via large-kernel attention graph convolutional network,” IEEE Transactions on Visualization and Computer Graphics , vol. 29, no. 5, pp. 2575–2585, 2023

    2023

  6. [11]

    Human3. 6m: Large scale datasets and predictive methods for 3d human sensing in natural environments,

    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

  7. [12]

    Exploiting tem- poral contexts with strided transformer for 3d human pose estimation,

    W. Li, H. Liu, R. Ding, M. Liu, P. Wang, and W. Yang, “Exploiting tem- poral contexts with strided transformer for 3d human pose estimation,” IEEE Transactions on Multimedia , vol. 25, pp. 1282–1293, 2022

    2022

  8. [13]

    Hourglass tokenizer for efficient transformer-based 3d human pose estimation,

    W. Li, M. Liu, H. Liu, P. Wang, J. Cai, and N. Sebe, “Hourglass tokenizer for efficient transformer-based 3d human pose estimation,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2024, pp. 604–613

    2024

  9. [14]

    Large separable kernel attention: Rethinking the large kernel attention design in cnn,

    K. W. Lau, L.-M. Po, and Y . A. U. Rehman, “Large separable kernel attention: Rethinking the large kernel attention design in cnn,” Expert Systems with Applications , vol. 236, p. 121352, 2024

    2024

  10. [15]

    Pelk: Parameter- efficient large kernel convnets with peripheral convolution,

    H. Chen, X. Chu, Y . Ren, X. Zhao, and K. Huang, “Pelk: Parameter- efficient large kernel convnets with peripheral convolution,” in Proceed- ings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2024, pp. 5557–5567. JOURNAL OF LATEX CLASS FILES, VOL. 14, ...

    2024

  11. [16]

    Mixste: Seq2seq mixed spatio-temporal encoder for 3d human pose estimation in video,

    J. Zhang, Z. Tu, J. Yang, Y . Chen, and J. Yuan, “Mixste: Seq2seq mixed spatio-temporal encoder for 3d human pose estimation in video,” in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2022, pp. 13 232–13 242

    2022

  12. [18]

    Motionbert: A unified perspective on learning human motion representations,

    W. Zhu, X. Ma, Z. Liu, L. Liu, W. Wu, and Y . Wang, “Motionbert: A unified perspective on learning human motion representations,” in Proceedings of the IEEE/CVF International Conference on Computer Vision, 2023, pp. 15 085–15 099

    2023

  13. [19]

    Transxnet: learning both global and local dynamics with a dual dynamic token mixer for visual recognition,

    M. Lou, H.-Y . Zhou, S. Yang, and Y . Yu, “Transxnet: learning both global and local dynamics with a dual dynamic token mixer for visual recognition,” arXiv preprint arXiv:2310.19380 , 2023

    2023 arXiv

  14. [20]

    Camera distance-aware top- down approach for 3d multi-person pose estimation from a single rgb image,

    G. Moon, J. Y . Chang, and K. M. Lee, “Camera distance-aware top- down approach for 3d multi-person pose estimation from a single rgb image,” in Proceedings of the IEEE/CVF international conference on computer vision, 2019, pp. 10 133–10 142

    2019

  15. [21]

    Integral human pose regression,

    X. Sun, B. Xiao, F. Wei, S. Liang, and Y . Wei, “Integral human pose regression,” in Proceedings of the European conference on computer vision (ECCV), 2018, pp. 529–545

    2018

  16. [22]

    Hemlets pose: Learning part-centric heatmap triplets for accurate 3d human pose estimation,

    K. Zhou, X. Han, N. Jiang, K. Jia, and J. Lu, “Hemlets pose: Learning part-centric heatmap triplets for accurate 3d human pose estimation,” in Proceedings of the IEEE/CVF international conference on computer vision, 2019, pp. 2344–2353

    2019

  17. [23]

    Probabilistic monocular 3d human pose estimation with normalizing flows,

    T. Wehrbein, M. Rudolph, B. Rosenhahn, and B. Wandt, “Probabilistic monocular 3d human pose estimation with normalizing flows,” in Proceedings of the IEEE/CVF international conference on computer vision, 2021, pp. 11 199–11 208

    2021

  18. [24]

    Ordinal depth supervision for 3d human pose estimation,

    G. Pavlakos, X. Zhou, and K. Daniilidis, “Ordinal depth supervision for 3d human pose estimation,” in Proceedings of the IEEE conference on computer vision and pattern recognition , 2018, pp. 7307–7316

    2018

  19. [25]

    Anatomy-aware 3d human pose estimation with bone-based pose decomposition,

    T. Chen, C. Fang, X. Shen, Y . Zhu, Z. Chen, and J. Luo, “Anatomy-aware 3d human pose estimation with bone-based pose decomposition,” IEEE Transactions on Circuits and Systems for Video Technology , vol. 32, no. 1, pp. 198–209, 2021

    2021

  20. [26]

    Attention mechanism exploits temporal contexts: Real-time 3d human pose reconstruction,

    R. Liu, J. Shen, H. Wang, C. Chen, S.-c. Cheung, and V . Asari, “Attention mechanism exploits temporal contexts: Real-time 3d human pose reconstruction,” in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , 2020, pp. 5064–5073

    2020

  21. [27]

    3d human pose estimation with spatial and temporal transformers,

    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, 2021, pp. 11 656–11 665

    2021

  22. [29]

    Cascaded pyramid network for multi-person pose estimation,

    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 , 2018, pp. 7103–7112

    2018

  23. [30]

    Stacked hourglass networks for human pose estimation,

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

    2016

  24. [31]

    Deep high-resolution repre- sentation learning for human pose estimation,

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

    2019

  25. [32]

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

    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, 2022, pp. 13 147–13 156

    2022

  26. [33]

    Diffusion-based 3d human pose estimation with multi- hypothesis aggregation,

    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,” arXiv preprint arXiv:2303.11579 , 2023

    2023 arXiv

  27. [34]

    Double-chain constraints for 3d human pose estimation in images and videos,

    H. Kang, Y . Wang, M. Liu, D. Wu, P. Liu, and W. Yang, “Double-chain constraints for 3d human pose estimation in images and videos,” arXiv preprint arXiv:2308.05298, 2023

    2023 arXiv

  28. [35]

    An image is worth 16x16 words: Transformers for image recognition at scale,

    D. Alexey, “An image is worth 16x16 words: Transformers for image recognition at scale,” arXiv preprint arXiv: 2010.11929 , 2020

    2010 arXiv

  29. [36]

    Swin transformer: Hierarchical vision transformer using shifted windows,

    Z. Liu, Y . Lin, Y . Cao, H. Hu, Y . Wei, Z. Zhang, S. Lin, and B. Guo, “Swin transformer: Hierarchical vision transformer using shifted windows,” in Proceedings of the IEEE/CVF international conference on computer vision, 2021, pp. 10 012–10 022

    2021

  30. [37]

    Understanding the effective receptive field in deep convolutional neural networks,

    W. Luo, Y . Li, R. Urtasun, and R. Zemel, “Understanding the effective receptive field in deep convolutional neural networks,” Advances in neural information processing systems , vol. 29, 2016

    2016

  31. [38]

    Vision transformers for dense prediction,

    R. Ranftl, A. Bochkovskiy, and V . Koltun, “Vision transformers for dense prediction,” in Proceedings of the IEEE/CVF international conference on computer vision , 2021, pp. 12 179–12 188

    2021

  32. [39]

    Contnet: Why not use convolution and transformer at the same time?

    H. Yan, Z. Li, W. Li, C. Wang, M. Wu, and C. Zhang, “Contnet: Why not use convolution and transformer at the same time?” arXiv preprint arXiv:2104.13497, 2021

    2021 arXiv

  33. [40]

    Scaling up your kernels to 31x31: Revisiting large kernel design in cnns,

    X. Ding, X. Zhang, J. Han, and G. Ding, “Scaling up your kernels to 31x31: Revisiting large kernel design in cnns,” in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , 2022, pp. 11 963–11 975

    2022

  34. [41]

    Visual attention network,

    M.-H. Guo, C.-Z. Lu, Z.-N. Liu, M.-M. Cheng, and S.-M. Hu, “Visual attention network,” Computational Visual Media, vol. 9, no. 4, pp. 733– 752, 2023

    2023

  35. [42]

    A convnet for the 2020s,

    Z. Liu, H. Mao, C.-Y . Wu, C. Feichtenhofer, T. Darrell, and S. Xie, “A convnet for the 2020s,” in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , 2022, pp. 11 976–11 986

    2022

  36. [43]

    Learning dynamic local context representations for infrared small target detec- tion,

    G. Zhang, G. Xu, H. Wang, S. Chen, Y . Shan, and X. Zhang, “Learning dynamic local context representations for infrared small target detec- tion,” 2024

    2024

  37. [44]

    Metaformer is actually what you need for vision,

    W. Yu, M. Luo, P. Zhou, C. Si, Y . Zhou, X. Wang, J. Feng, and S. Yan, “Metaformer is actually what you need for vision,” in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , 2022, pp. 10 819–10 829

    2022

  38. [45]

    Gaussian error linear units (gelus),

    D. Hendrycks and K. Gimpel, “Gaussian error linear units (gelus),” arXiv preprint arXiv:1606.08415, 2016

    2016 arXiv

  39. [46]

    Layer normalization,

    J. Lei Ba, J. R. Kiros, and G. E. Hinton, “Layer normalization,” ArXiv e-prints, pp. arXiv–1607, 2016

    2016

  40. [47]

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

    D. Mehta, H. Rhodin, D. Casas, P. Fua, O. Sotnychenko, W. Xu, and C. Theobalt, “Monocular 3d human pose estimation in the wild using improved cnn supervision,” in 2017 international conference on 3D vision (3DV). IEEE, 2017, pp. 506–516

    2017

  41. [48]

    P-stmo: Pre-trained spatial temporal many-to-one model for 3d human pose estimation,

    W. Shan, Z. Liu, X. Zhang, S. Wang, S. Ma, and W. Gao, “P-stmo: Pre-trained spatial temporal many-to-one model for 3d human pose estimation,” in European Conference on Computer Vision . Springer, 2022, pp. 461–478

    2022

  42. [49]

    Uncertainty-aware 3d human pose estimation from monocular video,

    J. Zhang, Y . Chen, and Z. Tu, “Uncertainty-aware 3d human pose estimation from monocular video,” in Proceedings of the 30th ACM International Conference on Multimedia , 2022, pp. 5102–5113

    2022

  43. [50]

    Gla- gcn: Global-local adaptive graph convolutional network for 3d human pose estimation from monocular video,

    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,” inProceedings of the IEEE/CVF International Conference on Computer Vision , 2023, pp. 8818–8829

    2023

  44. [52]

    Smpl: A skinned multi-person linear model,

    M. Loper, N. Mahmood, J. Romero, G. Pons-Moll, and M. J. Black, “Smpl: A skinned multi-person linear model,” in Seminal Graphics Papers: Pushing the Boundaries, Volume 2 , 2023, pp. 851–866

    2023

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.