Pith. sign in

REVIEW 5 major objections 6 minor 34 references

ProbPose: A Probabilistic Approach to 2D Human Pose Estimation

T0 review · 5 major / 6 minor · reviewed 2026-08-11 · deepseek-v4-flash

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

Pith's one-line read ProbPose treats each joint as a distribution, so it can say when a keypoint is outside the image, and that calibrated 'don't know' output improves localization both in and beyond the frame.

desk verdict The presence-probability and Ex-OKS ideas are useful, but the paper's headline out-of-image gains are really crop augmentation, and it needs a controlled baseline before those claims hold. read the letter →

arxiv 2412.02254 v1 pith:GK5SGMSW submitted 2024-12-03 cs.CV

Miroslav Purkrabek , Jiri Matas This is my paper · ORCID
classification cs.CV
keywords humanposeestimationprobabilitymapspresenceout-of-imagekeypointsCropCOCOEx-OKSmetriccalibrationcroppingaugmentation
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

ProbPose is a top-down 2D human pose estimator that replaces heatmaps with calibrated probability maps and adds a per-keypoint presence probability: the chance that the joint actually falls inside the model's activation window. The paper argues that current pose models ignore out-of-image keypoints during training and evaluation, so they guess when a joint is cropped out, and their confidence scores are uncalibrated for in/out decisions. Training with randomly cropped images lets the model localize keypoints beyond the image border, lifting CropCOCO mAP from 72.7 to 81.7 for a ViTPose-scale model while also slightly improving in-image accuracy on COCO (76.6 vs 75.9). The paper also introduces Ex-OKS, an extension of OKS that scores presence and location together, plus the CropCOCO dataset. A sympathetic reader should care because calibrated 'I cannot see this joint' output is what robotics, interaction, and safety-critical applications need, not just a point estimate.

What carries the argument

The central machinery is the probability map: a per-keypoint distribution over pixels in the activation window, normalized to sum to 1 by a Sparsemax activation, so its shape is learned rather than forced to a Gaussian. These maps are trained by an expected-risk version of the OKS loss, with a Sobel-gradient regularizer, and decoded by expected OKS maximization, which favors regions with large probability mass over sharp isolated peaks. Presence probability is a separate head trained with binary cross-entropy on keypoints displaced out of the window by random cropping; temperature scaling calibrates both outputs. A double-heatmap variant adds a second, wider activation window at the same output resolution, extending the model's reach to keypoints up to 25% beyond the bounding box, and the new CropCOCO dataset with the Ex-OKS metric supplies the training and evaluation protocol for out-of-image keypoints.

What would settle it

Build a test set of real images in which people are truncated by the frame or heavily occluded, manually annotate keypoints that lie outside the visible image or the activation window, and compare ProbPose's presence probability and out-of-image localization against the ViTPose baseline; if presence accuracy matches simple confidence thresholding or the out-of-image localization gain on this real data does not appear, the central transfer claim is falsified. A separate check: on those real truncated images, measure whether the calibrated top-x-percent probability-map bins actually contain x percent of ground-truth keypoints, since the probabilistic-query claim stands or falls with that calibration.

Watch

Extended reading notes

Core claim

On the paper's own terms, the central discovery is that keypoint localization and keypoint presence are two separate, jointly trainable quantities, and both belong in the output representation. ProbPose outputs a probability map over the activation window normalized to sum to 1, a presence probability that the keypoint is inside that window, a predicted OKS quality score, and a visibility flag; at inference it picks the location that maximizes expected OKS rather than the heatmap argmax. Trained with an OKS risk-minimization loss and a cropping augmentation that pushes keypoints outside the image, this representation reaches 81.7 mAP and 73.9 Ex-mAP on CropCOCO versus 72.7 and 66.5 for the ViTPose-s baseline, while slightly improving COCO in-image localization. The paper further shows that presence probability classifies in/out status more accurately than confidence thresholding, cutting the error by 45% on a balanced subset, and that calibrating the outputs makes probabilistic queries such as 'the smallest region containing this keypoint with 95% probability' meaningful.

Load-bearing premise

Randomly cropping COCO images produces a training and evaluation distribution that faithfully represents how keypoints end up outside real images or bounding boxes, and the Ex-OKS distance penalty reflects how much real applications penalize presence errors; if those do not hold, the headline gains on CropCOCO may not transfer to real-world out-of-image keypoints.

Editorial extensions

If this is right

  • Pose estimators can be trained to localize joints that leave the visible image: on CropCOCO the ProbPose-s model reaches 81.7 mAP and 73.9 Ex-mAP against 72.7 and 66.5 for ViTPose-s with the same backbone scale and ground-truth boxes.
  • In/out decisions should use a dedicated presence probability rather than the heatmap peak, since the paper reports a 45% error reduction on the balanced CropCOCO subset and 30% on the full set.
  • Calibrated probability maps turn pose outputs into probabilistic statements, such as a confidence region that contains the keypoint with a chosen probability, which raw heatmap confidence cannot provide.
  • The double-heatmap design buys a wider field of view without the resolution loss of enlarging the input, localizing keypoints up to 25% outside the bounding box with only a 0.4% AP drop on COCO.
  • Because Ex-OKS penalizes false positives and wrong presence calls, models trained or evaluated with it are discouraged from guessing joints that are not visible in the activation window.

Reading between the lines

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

  • Beyond the paper: the same presence-probability signal could serve as an explicit uncertainty input for downstream systems, letting a robot or controller distinguish 'the joint is visible but I am unsure exactly where' from 'the joint is not in view at all'.
  • Beyond the paper: because the paper treats the image border as one form of occlusion, the cropping augmentation could be extended to object occlusion or truncation, potentially improving crowded-scene and self-occlusion benchmarks.
  • Beyond the paper: Ex-OKS could be adopted as a plug-in protocol for bottom-up and single-stage pose estimators, whose activation window is the whole image, making presence evaluation a general standard rather than a crop-specific metric.
  • Beyond the paper: a natural next stress test is a dataset of organically truncated people, such as sports or surveillance footage, where keypoints leave the frame without synthetic preprocessing; that would test whether the crop-trained distribution transfers to real-world out-of-image keypoints.
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

5 major / 6 minor

Summary. ProbPose proposes a top-down 2D human pose estimation method that replaces heatmaps with normalized probability maps, trains them with an OKS-based expected-risk loss, and adds a predicted presence probability for each keypoint indicating whether it lies inside the activation window. The paper also introduces crop-based data augmentation to create out-of-image keypoints during training, a double-heatmap variant with a wider field of view, a new synthetic CropCOCO evaluation dataset, and an Extended OKS (Ex-OKS) metric that assigns penalties for presence/absence disagreements. Experiments on COCO, CropCOCO, and OCHuman report improved out-of-image localization on CropCOCO (mAP 81.7 vs. 72.7 for ViTPose-s), a small in-image COCO gain (76.6 vs. 75.9), and a 45% reduction in in/out classification error relative to thresholding ViTPose confidence.

Significance. The paper introduces several genuinely useful ideas: a principled probabilistic treatment of keypoint localization, a loss aligned with the OKS evaluation metric, a presence-probability output that is calibrated and distinct from confidence, and a new evaluation protocol for out-of-image keypoints. The release of code and models is a strength, and the ablation study honestly isolates the effect of crop augmentation from the probability-map representation. However, the central claim that the probabilistic representation improves out-of-image localization is not supported by the paper's own ablation, and the headline ex-image gains rest almost entirely on CropCOCO, a synthetic set generated with the same crop procedure used in training. The significance will be established only after the evaluation is strengthened with the missing control experiments and independent validation.

major comments (5)
  1. [Table 2] The ablation shows that crop augmentation alone yields CropCOCO mAP 81.7 and adding probability maps leaves it at 81.7, while without crop augmentation probability maps actually reduce CropCOCO mAP (72.4 vs. 73.7). Therefore the headline CropCOCO localization gain in Table 1 (81.7 vs. 72.7 for ViTPose-s) is attributable to the crop augmentation, not to the probabilistic representation. The manuscript should either revise the central claim that the probabilistic approach improves out-of-image localization or provide an experiment where probability maps improve out-of-image localization beyond the augmentation effect.
  2. [Sec. 5.1, Fig. 5] The claimed 45% in/out classification error reduction compares the presence probability of ProbPose (trained with crop augmentation) with the confidence of ViTPose (trained without crop augmentation). This conflates the effect of the presence head with the effect of the augmentation, because the augmentation provides the training signal that makes any in/out predictor learnable. A control with a ViTPose or ProbPose variant trained with the same crop augmentation but using heatmap confidence for the in/out decision is necessary to attribute the improvement to the presence probability.
  3. [Sec. 4.1] Ex-mAP is computed by selecting the optimal presence threshold for each model on the test set, as stated in the manuscript. This is a form of test-set threshold selection and can inflate the reported Ex-mAP gains, especially when the compared models have different confidence distributions. The authors should select thresholds on a validation set or report Ex-mAP as a function of threshold, rather than using the test-set optimum.
  4. [Table 1, OCHuman rows] On OCHuman, the standard ProbPose shows no Ex-mAP improvement over ViTPose (60.2 vs. 60.1), and the main gain appears only in the double-heatmap variant (61.2). Since OCHuman is the only out-of-domain dataset in the paper, this result weakens the claim that the probabilistic representation improves out-of-image and presence-aware performance in general; the claim currently rests almost entirely on CropCOCO, whose distribution is generated by the same crop procedure used for training. The authors should discuss this limitation explicitly and, if possible, evaluate on an independently created out-of-image benchmark.
  5. [Sec. 5.2 / Tables 1-2] The paper reports small differences, such as 0.7 mAP on COCO and 0.1 Ex-mAP on OCHuman, without error bars, multiple seeds, or statistical significance testing. Given that the central in-image improvement is only about 1%, the authors should provide variance estimates or repeated-run results to establish that the differences are not within training noise.
minor comments (6)
  1. [Throughout] The notation for the activation window is inconsistent: Eq. (1) and Eq. (2) use AW, while Eq. (5) and Appendix A use AM. Please unify.
  2. [Eq. (5)] The definition of de(AM, x') is not precise; it should state explicitly that the distance is measured from the activation-window boundary, not from some interior point, and should specify the behavior when the point lies inside the window.
  3. [Sec. 4.2] The description of CropCOCO does not specify the crop sampling distribution, the range of crop strengths, or how bounding boxes are recomputed; this information is needed to assess how representative the synthetic evaluation is and to reproduce the dataset.
  4. [Sec. 5.2, Table 2 caption] The table header says 'PropPose' while the paper uses 'ProbPose'; please correct the typo.
  5. [Sec. 3.1] The statement that 'probability maps always sum to 1' via Sparsemax is slightly misleading: Sparsemax maps produce nonnegative outputs that sum to 1 only over the entire activation window, but the paper later uses per-pixel losses; the relationship between the per-pixel interpretation and the global normalization should be clarified.
  6. [Appendix B] The calibration procedure is described as ensuring each quantile contains approximately 5% of ground-truth keypoints, but the precise binning, temperature-scaling fitting procedure, and whether calibration is performed on the training or validation split are not fully specified.

Circularity Check

1 steps flagged · score 2.0 of 10

Only mild circularity: the calibration temperature is fit on CropCOCO and then validated on CropCOCO; the core OKS-based localization derivation is independent.

  1. fitted input called prediction [Sec. 3.3 (Calibration); evaluation on CropCOCO in Secs. 4-5]
    "Probability maps are calibrated so that the top 5% of the map contains 5% of the keypoints, the top 10% contains 10% of the keypoints, and so on. We achieve this calibration through temperature scaling on CropCOCO. As a result, the calibrated probabilities align with the true underlying distribution, enabling more versatile evaluations beyond simple point estimates."

    The temperature-scaling parameter is fit to CropCOCO, and CropCOCO is then the dataset on which the calibrated probability maps and presence probabilities are evaluated (Tabs. 1-2, Fig. 5, Fig. 7). The calibration criterion stated in the same sentence (top 5% of the map contains 5% of the keypoints, etc.) is the very property the temperature fit enforces on CropCOCO. The reported agreement between predicted probabilities and empirical frequencies on CropCOCO is therefore a restatement of the fit objective rather than an independent predictive check. The claim that ProbPose 'predicts a calibrated probability' is thus supported on CropCOCO by construction, though it remains an independent claim on COCO and OCHuman.

full rationale

The paper's main derivation is self-contained: training probability maps with an OKS-based expected-risk loss and decoding by expected-OKS maximization is a deliberate alignment of the training objective with the evaluation metric, not an input-output equivalence. The COCO and OCHuman results provide external benchmarks, and Table 2 explicitly shows that crop augmentation, not the probability maps, drives the CropCOCO mAP gain (81.7 with and without probability maps), so the paper does not misattribute that localization gain. The introduced CropCOCO dataset and Ex-OKS metric define a new task rather than circularly deriving a result, although the CropCOCO evaluation distribution closely matches the training augmentation and the ViTPose baseline was not trained with crop augmentation; those are evaluation confounds, not circular reasoning. The only circularity I find is in the calibration claim: the temperature-scaling parameter is fit on CropCOCO (Sec. 3.3), after which CropCOCO is used to demonstrate calibrated probability maps and presence-probability improvements, making that portion of the evidence statistically forced. There is no load-bearing self-citation chain, and no uniqueness theorem or ansatz is imported from the authors' prior work. Overall, the central probabilistic-localization derivation has independent content, so the circularity score is low.

Assumptions & free parameters 4 free parameters · 5 assumptions · 3 invented entities

The central claims rest on two trained heads (localization probability map and presence probability), a smoothing regularizer, and a temperature scaling step; none of these are derived from first principles, and all involve fitted or hand-set hyperparameters. The free parameters and domain assumptions are not physical constants but design choices that materially affect the reported accuracy numbers, so they are listed explicitly.

free parameters (4)
  • Regularization weight alpha in LOKS loss = not reported
    Appears in Eq. 4; controls the balance between expected OKS risk and the Sobel-based smoothness prior; no value is stated in the text.
  • Temperature scaling T for probability map calibration = not reported
    Used in Section 3.3 on CropCOCO to make probability maps calibrated; the specific temperature value is not given.
  • Crop augmentation parameters = not reported
    Section 3.4 describes random cropping that pushes keypoints outside the image/activation window, but the strength and ranges of the crops are not specified.
  • Ex-mAP presence threshold per model = selected optimally on test set
    Section 4.1 states 'we select the optimal threshold for each model' when computing Ex-mAP, a post-hoc selection on the evaluation set that can inflate reported numbers.
assumptions (5)
  • domain assumption The image border behaves as a form of occlusion; cropped-out keypoints should be treated like occluded keypoints.
    Stated in Section 3.4 and Appendix A; this justifies using random cropping to generate training signal for presence probability.
  • domain assumption Random crop augmentation from COCO yields out-of-image keypoint samples representative of real deployment conditions.
    Section 3.4; the central out-of-image results depend on this transfer from synthetic crops to real-world scenarios, which is not established on an independent dataset.
  • domain assumption COCO annotations near bounding box borders are systematically biased inward.
    Appendix D; used to explain why ProbPose is penalized near borders and to motivate crop augmentation as a corrective.
  • domain assumption The posterior localization distribution is not Gaussian and should be learned rather than imposed.
    Section 3.1 motivates probability maps instead of fixed-Gaussian heatmaps; this is presented as a design choice with qualitative support, not a formal proof.
  • standard math Sparsemax provides calibrated probabilities after temperature scaling.
    Sparsemax (reference [19]) guarantees outputs sum to 1, but calibration is empirical and achieved via temperature scaling; the activation alone does not guarantee calibration.
invented entities (3)
  • Presence probability pp(kj) independent evidence
    purpose: Predicts whether keypoint j lies inside the activation window, replacing arbitrary confidence thresholding.
    Trained with binary cross-entropy and evaluated against CropCOCO labels; it is a new output variable with a directly testable falsifiable evaluation.
  • Ex-OKS metric independent evidence
    purpose: Extends OKS to out-of-window keypoints by penalizing presence misclassification with distance from the activation window edge.
    Can be computed for any model's outputs and is publicly defined, but its distance-based penalty is an unvalidated modeling choice that affects reported improvements.
  • CropCOCO dataset independent evidence
    purpose: Evaluation benchmark with out-of-image keypoints for testing presence probability and out-of-image localization.
    Constructed from COCO with public source imagery, making it reproducible; whether it reflects real-world domain shift is unresolved.

how reviews work

0 comments
Cite this review

Pith. "Pith review of ProbPose: A Probabilistic Approach to 2D Human Pose Estimation." pith.science (2026). https://pith.science/paper/GK5SGMSW

@misc{pith2026241202254,
  author       = {Pith},
  title        = {Pith review of: ProbPose: A Probabilistic Approach to 2D Human Pose Estimation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/GK5SGMSW}},
  note         = {Machine review of arXiv:2412.02254}
}
read the original abstract

Current Human Pose Estimation methods have achieved significant improvements. However, state-of-the-art models ignore out-of-image keypoints and use uncalibrated heatmaps as keypoint location representations. To address these limitations, we propose ProbPose, which predicts for each keypoint: a calibrated probability of keypoint presence at each location in the activation window, the probability of being outside of it, and its predicted visibility. To address the lack of evaluation protocols for out-of-image keypoints, we introduce the CropCOCO dataset and the Extended OKS (Ex-OKS) metric, which extends OKS to out-of-image points. Tested on COCO, CropCOCO, and OCHuman, ProbPose shows significant gains in out-of-image keypoint localization while also improving in-image localization through data augmentation. Additionally, the model improves robustness along the edges of the bounding box and offers better flexibility in keypoint evaluation. The code and models are available on https://mirapurkrabek.github.io/ProbPose/ for research purposes.

Figures

Figures reproduced from arXiv: 2412.02254 by the authors.

Figure 1
Figure 1. ProbPose (left) vs. ViTPose [31] (right) output on a cropped image with the dark part removed. The bounding box, on which the pose estimators operate, is detected by YOLOX-x [5]. ProbPose estimates keypoints outside of the cropped image much better, including the right leg which is wrongly aligned with the left leg by ViTPose. through an appropriate choice of representation and loss function. Top-down methods locali… view at source ↗
Figure 2
Figure 2. COCO (first row) and the proposed CropCOCO (second [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. UDP decoding vs. expected OKS maximization. In the [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figures from the paper (6 more)
Figure 4
Figure 4. Figure 4: Visualizations of top-down approaches. Rectangles [PITH_FULL_IMAGE:figures/full_fig_p005_4.png]
Figure 5
Figure 5. Figure 5: Keypoint presence in the activation window. Compari [PITH_FULL_IMAGE:figures/full_fig_p007_5.png]
Figure 6
Figure 6. Figure 6: A scheme explaining where keypoints could be in the im [PITH_FULL_IMAGE:figures/full_fig_p011_6.png]
Figure 7
Figure 7. Figure 7: Calibration curves of probability maps (a) before and af [PITH_FULL_IMAGE:figures/full_fig_p012_7.png]
Figure 8
Figure 8. Figure 8: Decoding predicted probability maps throught UDP [PITH_FULL_IMAGE:figures/full_fig_p013_8.png]
Figure 10
Figure 10. Figure 10: Ground truth annotation (left) vs. ProbPose-s (right) on [PITH_FULL_IMAGE:figures/full_fig_p014_10.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

34 extracted references · 28 canonical work pages

  1. [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 (CVPR), 2014. 3

    work page 2014

  2. [2]

    Adversarial semantic data augmentation for hu- man pose estimation

    Yanrui Bin, Xuan Cao, Xinya Chen, Yanhao Ge, Ying Tai, Chengjie Wang, Jilin Li, Feiyue Huang, Changxin Gao, and Nong Sang. Adversarial semantic data augmentation for hu- man pose estimation. In European Conference on Computer Vision, 2020. 12

    work page 2020

  3. [3]

    Realtime multi-person 2d pose estimation using part affinity fields

    Zhe Cao, Tomas Simon, Shih-En Wei, and Yaser Sheikh. Realtime multi-person 2d pose estimation using part affinity fields. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 7291–7299, 2017. 2

    2017

  4. [4]

    Wing loss for robust facial landmark localisation with convolutional neural networks

    Zhen-Hua Feng, Josef Kittler, Muhammad Awais, Patrik Hu- ber, and Xiao-Jun Wu. Wing loss for robust facial landmark localisation with convolutional neural networks. InProceed- ings of the IEEE conference on computer vision and pattern recognition, pages 2235–2245, 2018. 3

    work page 2018

  5. [5]

    Yolox: Exceeding yolo series in 2021

    Z Ge. Yolox: Exceeding yolo series in 2021. arXiv preprint arXiv:2107.08430, 2021. 1

    arXiv 2021

  6. [6]

    Bottom-up human pose estimation via disentan- gled keypoint regression

    Zigang Geng, Ke Sun, Bin Xiao, Zhaoxiang Zhang, and Jing- dong Wang. Bottom-up human pose estimation via disentan- gled keypoint regression. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages 14676–14686, 2021. 2

    work page 2021

  7. [7]

    On the Calibration of Human Pose Estimation

    Kerui Gu, Rongyu Chen, and Angela Yao. On the calibration of human pose estimation. arXiv preprint arXiv:2311.17105,

    work page Pith review arXiv

  8. [8]

    The devil is in the details: Delving into unbiased data processing for human pose estimation

    Junjie Huang, Zheng Zhu, Feng Guo, and Guan Huang. The devil is in the details: Delving into unbiased data processing for human pose estimation. In The IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2020. 2, 12

    work page 2020

Show all 34 references
  1. [9]

    Aid: Pushing the performance boundary of human pose es- timation with information dropping augmentation

    Junjie Huang, Zheng Zhu, Guan Huang, and Dalong Du. Aid: Pushing the performance boundary of human pose es- timation with information dropping augmentation. arXiv preprint arXiv:2008.07139, 2020. 2, 3, 6, 8

    2008 arXiv

  2. [10]

    Cornernet: Detecting objects as paired keypoints

    Hei Law and Jia Deng. Cornernet: Detecting objects as paired keypoints. In Proceedings of the European confer- ence on computer vision (ECCV), pages 734–750, 2018. 3

    2018

  3. [11]

    Crowdpose: Efficient crowded scenes pose estimation and a new benchmark

    Jiefeng Li, Can Wang, Hao Zhu, Yihuan Mao, Hao-Shu Fang, and Cewu Lu. Crowdpose: Efficient crowded scenes pose estimation and a new benchmark. arXiv preprint arXiv:1812.00324, 2018. 3

    2018 arXiv

  4. [12]

    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. 3

    2021

  5. [13]

    Belongie, James Hays, Pietro Perona, Deva Ramanan, Piotr Doll ´ar, and C

    Tsung-Yi Lin, Michael Maire, Serge J. Belongie, James Hays, Pietro Perona, Deva Ramanan, Piotr Doll ´ar, and C. Lawrence Zitnick. Microsoft coco: Common objects in context. In European Conference on Computer Vision, 2014. 3

    2014

  6. [14]

    Focal loss for dense object detection

    Tsung-Yi Lin, Priya Goyal, Ross Girshick, Kaiming He, and Piotr Doll´ar. Focal loss for dense object detection. In Pro- ceedings of the IEEE international conference on computer vision, pages 2980–2988, 2017. 3

    2017

  7. [15]

    Swin transformer: Hierarchical vision transformer using shifted windows

    Ze Liu, Yutong Lin, Yue Cao, Han Hu, Yixuan Wei, Zheng Zhang, Stephen Lin, and Baining Guo. Swin transformer: Hierarchical vision transformer using shifted windows. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 10012–10022, 2021. 2, 7

    2021

  8. [16]

    Rtmo: Towards high-performance one- stage real-time multi-person pose estimation

    Peng Lu, Tao Jiang, Yining Li, Xiangtai Li, Kai Chen, and Wenming Yang. Rtmo: Towards high-performance one- stage real-time multi-person pose estimation. In Proceed- ings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 1491–1500, 2024. 3

    2024

  9. [17]

    Human pose regression by combining indirect part detection and contextual information

    Diogo C Luvizon, Hedi Tabia, and David Picard. Human pose regression by combining indirect part detection and contextual information. Computers & Graphics, 85:15–22,

  10. [18]

    Yolo-pose: Enhancing yolo for multi person pose estimation using object keypoint similarity loss

    Debapriya Maji, Soyeb Nagori, Manu Mathew, and Deepak Poddar. Yolo-pose: Enhancing yolo for multi person pose estimation using object keypoint similarity loss. In Proceed- ings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 2637–2646, 2022. 3, 4

    2022

  11. [19]

    From softmax to sparsemax: A sparse model of attention and multi-label clas- sification

    Andre Martins and Ramon Astudillo. From softmax to sparsemax: A sparse model of attention and multi-label clas- sification. In International conference on machine learning, pages 1614–1623. PMLR, 2016. 4

    2016

  12. [20]

    Heatmap distribution matching for human pose estima- tion

    Haoxuan Qu, Li Xu, Yujun Cai, Lin Geng Foo, and Jun Liu. Heatmap distribution matching for human pose estima- tion. Advances in Neural Information Processing Systems , 35:24327–24339, 2022. 3

    2022

  13. [21]

    End-to-end multi-person pose estimation with transformers

    Dahu Shi, Xing Wei, Liangqi Li, Ye Ren, and Wenming Tan. End-to-end multi-person pose estimation with transformers. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 11069–11078, 2022. 2

    2022

  14. [22]

    End-to- end trainable multi-instance pose estimation with transform- ers

    Lucas Stoffl, Maxime Vidal, and Alexander Mathis. End-to- end trainable multi-instance pose estimation with transform- ers. arXiv preprint arXiv:2103.12115, 2021. 2

    2021 arXiv

  15. [23]

    Deep high-resolution representation learning for human pose esti- mation

    Ke Sun, Bin Xiao, Dong Liu, and Jingdong Wang. Deep high-resolution representation learning for human pose esti- mation. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 5693–5703, 2019. 2

    2019

  16. [24]

    Compositional human pose regression

    Xiao Sun, Jiaxiang Shang, Shuang Liang, and Yichen Wei. Compositional human pose regression. In Proceedings of the IEEE international conference on computer vision, pages 2602–2611, 2017. 2

    2017

  17. [25]

    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 (ECCV), pages 529–545, 2018. 2

    2018

  18. [26]

    Deeppose: Human pose estimation via deep neural networks

    Alexander Toshev and Christian Szegedy. Deeppose: Human pose estimation via deep neural networks. 2014 IEEE Con- ference on Computer Vision and Pattern Recognition, pages 1653–1660, 2013. 2

    2014

  19. [27]

    Contextual instance de- coupling for robust multi-person pose estimation

    Dongkai Wang and Shiliang Zhang. Contextual instance de- coupling for robust multi-person pose estimation. In Pro- ceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 11060–11068, 2022. 2 9

    2022

  20. [28]

    Pvt v2: Improved baselines with pyramid vision transformer

    Wenhai Wang, Enze Xie, Xiang Li, Deng-Ping Fan, Kaitao Song, Ding Liang, Tong Lu, Ping Luo, and Ling Shao. Pvt v2: Improved baselines with pyramid vision transformer. Computational Visual Media, 8(3):415–424, 2022. 2, 7

    2022

  21. [29]

    Adaptive wing loss for robust face alignment via heatmap regression

    Xinyao Wang, Liefeng Bo, and Li Fuxin. Adaptive wing loss for robust face alignment via heatmap regression. In Proceedings of the IEEE/CVF international conference on computer vision, pages 6971–6981, 2019. 3

    2019

  22. [30]

    Adaptivepose++: A powerful single-stage network for multi-person pose re- gression

    Yabo Xiao, Xiaojuan Wang, Dongdong Yu, Kai Su, Lei Jin, Mei Song, Shuicheng Yan, and Jian Zhao. Adaptivepose++: A powerful single-stage network for multi-person pose re- gression. arXiv preprint arXiv:2210.04014, 2022. 2

    2022 arXiv

  23. [31]

    ViTPose: Simple vision transformer baselines for human pose estimation

    Yufei Xu, Jing Zhang, Qiming Zhang, and Dacheng Tao. ViTPose: Simple vision transformer baselines for human pose estimation. In Advances in Neural Information Pro- cessing Systems, 2022. 1, 2, 7

    2022

  24. [32]

    Hrformer: High- resolution vision transformer for dense predict

    Yuhui Yuan, Rao Fu, Lang Huang, Weihong Lin, Chao Zhang, Xilin Chen, and Jingdong Wang. Hrformer: High- resolution vision transformer for dense predict. Advances in neural information processing systems , 34:7281–7293,

  25. [33]

    Rosin, Zixi Cai, Han Xi, Dingcheng Yang, Hao-Zhi Huang, and Shi-Min Hu

    Song-Hai Zhang, Ruilong Li, Xin Dong, Paul L. Rosin, Zixi Cai, Han Xi, Dingcheng Yang, Hao-Zhi Huang, and Shi-Min Hu. Pose2seg: Detection free human instance segmentation,

  26. [2019]

    A scheme explaining where keypoints could be in the im- age

    3 10 ProbPose: A Probabilistic Approach to 2D Human Pose Estimation Supplementary Material AB CE D Figure 6. A scheme explaining where keypoints could be in the im- age. Rectangles represent the bounding box , the model input and the activation window (usually coincides with m...

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.