Pith. sign in

REVIEW 4 major objections 6 minor 25 references

SimpleDepthPose: Fast and Reliable Human Pose Estimation with RGBD-Images

T0 review · 4 major / 6 minor · reviewed 2026-08-09 · deepseek-v4-flash

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

Pith's one-line read SimpleDepthPose estimates 3D human poses by reading depth values at visible 2D joints and averaging them, with no learned triangulation and no additional training.

desk verdict Simple, training-free RGBD pose estimator with plausible gains on occluded scenes; headline comparison rests on inherited baseline numbers, but the method itself is honest and reproducible. read the letter →

arxiv 2501.18478 v2 pith:SYUD4CE7 submitted 2025-01-30 cs.CV

Daniel Bermuth , Alexander Poeppel , Wolfgang Reif This is my paper · ORCID
classification cs.CV
keywords multi-viewposeestimationRGBDimagesdepth-based3Dmulti-personno-traininggeneralizationoutlierfilteringPCPmetricdepthextraction
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

This paper introduces SimpleDepthPose, a method for estimating the 3D poses of multiple people from several RGBD cameras. It claims that reliable multi-view pose estimation can be achieved without learned triangulation or any additional training: take off-the-shelf 2D joint detections from each color image, read the depth at each visible joint, and fuse the resulting 3D proposals by grouping, outlier filtering, and averaging. On the MVOR operating-room dataset it reports a PCP (percentage of correct parts) of 74.0, recall of 96.6, and F1 of 85.4, higher than all compared baselines, while running at 37.2 frames per second. The value of the claim is practical: safety-critical applications such as human-robot collaboration need fast, reliable pose estimates, and depth cameras are increasingly available.

What carries the argument

The central mechanism is a pipeline of five simple steps: 2D pose prediction per view, depth extraction, person grouping, outlier filtering, and fusion by averaging. The depth extraction uses a cross-shaped set of pixels around each detected joint and takes the median value, which suppresses outliers at object edges; the per-joint offsets account for the fact that a depth camera measures the surface while the target is the joint center. The outlier filter checks each joint proposal against the averaged center of its neighboring joints and discards proposals with implausibly long limbs. The grouping step assigns each 3D proposal to the closest person from the previous time step, creating new persons when needed. Together these steps let the method skip the learned voxel or triangulation networks used by other approaches.

What would settle it

Run SimpleDepthPose on a multi-view RGBD sequence using a 2D pose estimator that does not suppress occluded keypoints, such as the default COCO-trained HigherHRNet with its refinement step kept, and compare the F1 and MPJPE to the same pipeline with visibility-finetuned keypoints; a sharp rise in invalid predictions and joint errors would confirm that the visible-only assumption carries the result. Alternatively, test on subjects wearing thick or loose clothing and measure the systematic offset in joint positions.

Watch

Extended reading notes

Core claim

On the paper's own terms, the central discovery is that depth information can replace the learnable triangulation step entirely. For each color image the method predicts only directly visible 2D keypoints, reads the median depth from a cross-shaped window around each keypoint, adds a static per-joint offset (for example 3 cm for shoulders and knees, 1 cm for wrists) to reach the joint center, and transforms the resulting camera-space points into world coordinates. Person proposals are grouped across views and over time, outliers whose limb lengths exceed a threshold are discarded, and the remaining proposals are averaged. The result is a detection rate and accuracy that the paper reports as above all baselines on the MVOR dataset and comparable to learned methods on the Panoptic dataset, at 37.2 FPS on a single Nvidia-3090. In strongly occluded settings, a single view with depth is enough to detect a person, which triangulation-based methods cannot do.

Load-bearing premise

The method assumes the 2D pose estimator outputs only directly visible joints and never occluded ones, because an occluded joint has no meaningful depth value; if this fails, wrong 3D joints and reduced detection accuracy follow.

Editorial extensions

If this is right

  • If the reported results hold, depth sensors can replace learned triangulation networks, removing the need for per-scene training data in multi-view pose estimation.
  • In occluded scenes, a person visible in only one camera can still be detected and posed, because depth provides the third dimension from a single view.
  • The method runs at 37.2 FPS on an Nvidia-3090, making it suitable for real-time and safety-critical applications such as human-robot collaboration.
  • Because the 3D pipeline is training-free, it can be quickly adapted to other keypoint sets, as demonstrated with whole-body keypoints.
  • The ablation of camera count shows performance degrades gracefully with fewer cameras, so the approach can be tuned to available hardware.

Reading between the lines

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

  • An implicit consequence is that the method's accuracy is capped by the depth sensor's noise and calibration; future work could couple it with a lightweight learned refinement of the joint offsets rather than static proportions.
  • The visibility-only requirement suggests a natural test for 2D detectors: the same pipeline could serve as a probe for whether a detector suppresses occluded keypoints, since occluded outputs will show up as large limb-length outliers.
  • The point-cloud fusion experiments indicate the method can consume depth from non-camera sources, which would matter for robotics setups that already produce fused point clouds.
  • A temporal extension with a simple Kalman filter, avoided here to prevent lag, could be reintroduced selectively for slow-moving joints to improve accuracy without hurting fast motions.
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

4 major / 6 minor

Summary. The paper presents SimpleDepthPose, a training-free algorithm for multi-view, multi-person 3D pose estimation from RGBD images. The method uses an off-the-shelf 2D pose estimator, extracts depth values from aligned depth images for each visible joint, groups per-view 3D proposals into persons across time, filters outliers with limb-length heuristics, and averages the remaining proposals. The authors evaluate the method on the MVOR and Panoptic datasets, reporting strong detection rates especially under occlusions, competitive accuracy, and high frame rates. They also provide ablations on camera count, visibility finetuning, joint offsets, and alternative point-cloud fusion strategies. The central claim is that SimpleDepthPose generalizes better than prior methods without any additional training, while being fast and reliable.

Significance. If the reported results are independently reproduced, this is a practically valuable contribution: it is a simple, deterministic, training-free baseline that requires no learned 3D fusion and runs at real-time speed. The public release of code is a concrete strength, as is the fact that the method fits no parameters to the evaluation data, reducing circularity concerns. The paper's ablations are informative, particularly the camera-count study and the analysis of visibility finetuning. However, the comparative claim of 'best generalization' rests on baseline numbers taken mostly from the authors' own prior paper, with no error bars or protocol-equivalence evidence; the method also depends on load-bearing assumptions (visible-joint-only 2D detection, static per-joint offsets) that are not quantitatively stress-tested. The contribution is therefore plausible but not yet fully established.

major comments (4)
  1. [Table 1 and Table 2] The baseline numbers without extra citations are taken from the authors' own prior work [1], as stated in the table captions. Since the paper's headline conclusion is explicitly comparative ('shows the best generalization results among other methods'), the absence of independent re-runs or a detailed equivalence argument is load-bearing. The reported 20-point PCP lead over VoxelKeypointFusion on MVOR could shrink if the baselines were evaluated under slightly different keypoint filtering, person-visibility criteria, or depth alignment. Please provide either a same-protocol re-run of all baselines, or a precise specification of what was shared (2D detections, evaluation script, thresholds) and an analysis of how protocol choices affect the comparison. Reporting variances or error bars would additionally allow an assessment of significance.
  2. [Section 3, Step (1)] The method explicitly requires that the 2D pose estimator predict only directly visible keypoints, because occluded joints produce meaningless depth values. This is a load-bearing assumption, yet the paper offers no quantitative evidence of how often the visibility-finetuned HigherHRNet model emits occluded joints on MVOR or Panoptic, nor how the outlier filter behaves when this assumption fails. Please report the precision of the visible-joint prediction (e.g., the fraction of emitted keypoints that are actually visible in the depth image) and include an ablation with a standard COCO detector without visibility finetuning, so that the sensitivity of the downstream method to this requirement is quantified.
  3. [Section 3, Step (2)] The static per-joint depth offsets (e.g., 3 cm for shoulders/knees, 1 cm for wrists) are estimated from 'normal human proportions' and can be adapted, but no sensitivity analysis is provided. Since the method is positioned as reliable in varied settings, including thick clothing or unusual body proportions, the impact of these offsets on MPJPE and PCP should be quantified. A simple ablation varying the offsets (or a discussion of the expected error range) would indicate how much the reported accuracy depends on these hand-set values.
  4. [Tables 1–3] No error bars, confidence intervals, or repeated-run variances are reported for any metric. Because the central claims are comparative (higher recall, higher F1, faster than other methods), single-number entries without variance information make it impossible to judge whether differences are meaningful. Please add variance information over evaluation videos or subjects, or at least state the number of frames used for each metric and dataset.
minor comments (6)
  1. [Section 3, Step (2)] The cross-shape depth extraction is described textually and in Figure 2, but the dimensions of the cross (pixel sizes, arm lengths) are never specified; please give concrete numbers.
  2. [Section 4] 'operation room' should be 'operating room' for consistency with standard terminology.
  3. [Abstract and Section 6] The abstract claims the method is 'adaptable to different keypoints', but Section 6 shows that the visibility finetuning cannot be applied to face and finger keypoints, and whole-body performance is limited. Please temper the claim or report quantitative whole-body results.
  4. [Tables 1 and 2] The two VoxelKeypointFusion rows (one without depth, one with depth) are easy to confuse; consider labeling the rows explicitly as 'RGB' and 'RGBD'.
  5. [Throughout] The name 'HigherHrNet' is written with inconsistent capitalization; use 'HigherHRNet' consistently.
  6. [Section 1] The source-code URL is a bare repository link; please provide a specific commit or version for reproducibility.

Circularity Check

1 steps flagged · score 4.0 of 10

Comparative 'best generalization' claim rests on baseline numbers taken from the authors' own prior paper [1] without independent reproduction; the method itself is not fitted to the evaluation data.

  1. self citation load bearing [Table 1 caption (Section 4); Section 7 Conclusion]
    "All other results without extra citations are taken from [1]. ... This paper showed, through an evaluation of different datasets, that the proposed SimpleDepthPose algorithm is a very fast and reliable approach if depth data is available, and also shows the best generalization results among other methods, without requiring any additional training."

    The central comparative claim ('best generalization results among other methods') is established only by comparing Table 1/2 numbers. The table caption states that all non-cited baseline results are taken from [1], which is the authors' own prior VoxelKeypointFusion paper. Thus the baseline competitor scores (e.g., VoxelKeypointFusion PCP 54.0, OpenPTrack 11.7 on MVOR) are self-reported in the authors' earlier work and are not re-run or independently verified in the present paper. The evaluation metrics are also defined in [1]. The superiority conclusion therefore reduces to a self-citation chain: [1] supplies the baseline values and the evaluation protocol, and the present paper declares itself better than them.

full rationale

The SimpleDepthPose method is a hand-designed pipeline (2D visible-joint detection, median depth extraction with fixed per-joint offsets from human proportions, temporal grouping, outlier filter, averaging). No parameter is fitted to the evaluation datasets, so there is no fitted-input-called-prediction circularity. The per-joint offsets are derived from general human proportions, not from MVOR or Panoptic labels. The ablation studies and whole-body extension do not hide a reduction of a claimed prediction to an input. The only load-bearing circularity is in the comparative evaluation: the headline 'best generalization results' relies on baseline numbers taken from the authors' own prior paper [1] without independent reproduction. This makes the central comparison partly self-referential, but the method itself has independent algorithmic content, so the score is 4 rather than higher.

Assumptions & free parameters 6 free parameters · 4 assumptions · 0 invented entities

The method introduces no learned parameters and no new physical entities. Its performance rests on a set of hand-chosen thresholds and offsets, and on the domain assumptions that the depth images are calibrated and aligned, that the 2D detector can suppress occluded joints, and that human limb proportions are roughly standard. The evaluation also depends on baseline numbers taken largely from the authors' own VoxelKeypointFusion paper (Tables 1 and 2).

free parameters (6)
  • per-joint depth offset = 3 cm for shoulders/knees, 1 cm for wrists (hand-chosen)
    Added to the depth reading to correct for the sensor measuring body surface instead of joint center (Section 3, step 2). Ablation shows it improves mean error.
  • cross-shape window size for median depth = not stated in text (set in code)
    Defines the neighborhood around each 2D joint used to compute the median depth; affects robustness to depth edges.
  • outlier filter threshold = 0.5 m default
    Maximum allowed distance from a joint to the averaged center of its neighbors; used to reject impossible limb lengths (Section 3, step 4).
  • topk nearest proposals to average = 3
    Number of closest proposals used to compute the final joint position (Section 3, step 5).
  • person matching distance threshold and frame drop count = not stated in text (set in code)
    Thresholds for assigning 3D proposals to persons from the previous time step and for deleting unmatched persons (Section 3, step 3).
  • voxel resolution for point-cloud fusion ablation = 5 cm
    Used only in the pc2vmap ablation variant (Table 3).
assumptions (4)
  • domain assumption Color and depth images are aligned and camera extrinsics are calibrated.
    The method reads depth at 2D joint coordinates and transforms to world coordinates, so alignment and calibration are assumed (Section 3, steps 2-3).
  • domain assumption The depth sensor measures the distance to the body surface, so a fixed offset is needed to approximate joint centers.
    Justifies the per-joint offsets; assumes normal human proportions (Section 3, step 2).
  • ad hoc to paper A 2D pose estimator can be finetuned to output only visible joints reliably.
    The authors finetune HigherHRNet on COCO to predict only visible joints; if this is not achievable for other keypoint types (e.g., fingers), performance degrades (Sections 3 and 6).
  • ad hoc to paper Depth errors are mostly far outliers that can be removed by limb-length and neighbor filters.
    The outlier filter assumes wrong proposals are far from the true joint, which is the basis for steps 4 and 5 (Section 3).

how reviews work

0 comments
Cite this review

Pith. "Pith review of SimpleDepthPose: Fast and Reliable Human Pose Estimation with RGBD-Images." pith.science (2026). https://pith.science/paper/SYUD4CE7

@misc{pith2026250118478,
  author       = {Pith},
  title        = {Pith review of: SimpleDepthPose: Fast and Reliable Human Pose Estimation with RGBD-Images},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/SYUD4CE7}},
  note         = {Machine review of arXiv:2501.18478}
}
read the original abstract

In the rapidly advancing domain of computer vision, accurately estimating the poses of multiple individuals from various viewpoints remains a significant challenge, especially when reliability is a key requirement. This paper introduces a novel algorithm that excels in multi-view, multi-person pose estimation by incorporating depth information. An extensive evaluation demonstrates that the proposed algorithm not only generalizes well to unseen datasets, and shows a fast runtime performance, but also is adaptable to different keypoints. To support further research, all of the work is publicly accessible.

Figures

Figures reproduced from arXiv: 2501.18478 by the authors.

Figure 1
Figure 1. Example of a multi-person pose estimation from multi [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Visualization of the cross-shape used to extract the depth [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Example of the proposals for each view with some joint [PITH_FULL_IMAGE:figures/full_fig_p003_3.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

25 extracted references · 19 canonical work pages

  1. [1]

    V oxelKeypointFusion: Generalizable Multi-View Multi- Person Pose Estimation

    Daniel Bermuth, Alexander Poeppel, and Wolfgang Reif. V oxelKeypointFusion: Generalizable Multi-View Multi- Person Pose Estimation. arXiv preprint arXiv:2410.18723,

    arXiv

  2. [2]

    Real-time marker-less multi-person 3D pose es- timation in RGB-depth camera networks

    Marco Carraro, Matteo Munaro, Jeff Burke, and Emanuele Menegatti. Real-time marker-less multi-person 3D pose es- timation in RGB-depth camera networks. In Intelligent Au- tonomous Systems 15: Proceedings of the 15th International Conference IAS-15, pages 534–545. Springer, 2019. 2

    work page 2019

  3. [3]

    Higherhrnet: Scale- aware representation learning for bottom-up human pose es- timation

    Bowen Cheng, Bin Xiao, Jingdong Wang, Honghui Shi, Thomas S Huang, and Lei Zhang. Higherhrnet: Scale- aware representation learning for bottom-up human pose es- timation. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 5386–5395,

    work page

  4. [4]

    TEMPO: Efficient multi-view pose estimation, tracking, and forecasting

    Rohan Choudhury, Kris M Kitani, and L ´aszl´o A Jeni. TEMPO: Efficient multi-view pose estimation, tracking, and forecasting. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 14750–14760, 2023. 2

    work page 2023

  5. [5]

    Part-aware measurement for robust multi-view multi-human 3d pose estimation and track- ing

    Hau Chu, Jia-Hong Lee, Yao-Chih Lee, Ching-Hsien Hsu, Jia-Da Li, and Chu-Song Chen. Part-aware measurement for robust multi-view multi-human 3d pose estimation and track- ing. In Proceedings of the IEEE/CVF Conference on Com- puter Vision and Pattern Recognition , pages 1472–1481,

  6. [6]

    Fast and Robust Multi-Person 3D Pose Esti- mation from Multiple Views

    Junting Dong, Wen Jiang, Qixing Huang, Hujun Bao, and Xiaowei Zhou. Fast and Robust Multi-Person 3D Pose Esti- mation from Multiple Views. 2019. 2

    work page 2019

  7. [7]

    Fusing information from multiple 2D depth cam- eras for 3D human pose estimation in the operating room

    Lasse Hansen, Marlin Siebert, Jasper Diesel, and Mattias P Heinrich. Fusing information from multiple 2D depth cam- eras for 3D human pose estimation in the operating room. International journal of computer assisted radiology and surgery, 14:1871–1879, 2019. 2

    work page 2019

  8. [8]

    Learnable triangulation of human pose

    Karim Iskakov, Egor Burkov, Victor Lempitsky, and Yury Malkov. Learnable triangulation of human pose. InProceed- ings of the IEEE/CVF international conference on computer vision, pages 7718–7727, 2019. 2

    work page 2019

Show all 25 references
  1. [9]

    RTMPose: Real- Time Multi-Person Pose Estimation based on MMPose

    Tao Jiang, Peng Lu, Li Zhang, Ningsheng Ma, Rui Han, Chengqi Lyu, Yining Li, and Kai Chen. RTMPose: Real- Time Multi-Person Pose Estimation based on MMPose. arXiv preprint arXiv:2303.07399, 2023. 5

    2023 arXiv

  2. [10]

    Panoptic studio: A massively multiview system for social motion capture

    Hanbyul Joo, Hao Liu, Lei Tan, Lin Gui, Bart Nabbe, Iain Matthews, Takeo Kanade, Shohei Nobuhara, and Yaser Sheikh. Panoptic studio: A massively multiview system for social motion capture. In Proceedings of the IEEE Inter- national Conference on Computer Vision, pages 3334–3342,

  3. [11]

    A generalizable approach for multi-view 3d human pose re- gression

    Abdolrahim Kadkhodamohammadi and Nicolas Padoy. A generalizable approach for multi-view 3d human pose re- gression. Machine Vision and Applications , 32(1):6, 2021. 4

    2021

  4. [12]

    A multi-view RGB-D ap- proach for human pose estimation in operating rooms

    Abdolrahim Kadkhodamohammadi, Afshin Gangi, Michel de Mathelin, and Nicolas Padoy. A multi-view RGB-D ap- proach for human pose estimation in operating rooms. In 2017 IEEE winter conference on applications of computer vision (WACV), pages 363–372. IEEE, 2017. 2, 4

    2017

  5. [13]

    Microsoft coco: Common objects in context

    Tsung-Yi Lin, Michael Maire, Serge Belongie, James Hays, Pietro Perona, Deva Ramanan, Piotr Doll´ar, and C Lawrence Zitnick. Microsoft coco: Common objects in context. In Computer Vision–ECCV 2014: 13th European Conference, Zurich, Switzerland, September 6-12, 2014, Proceeding...

    2014

  6. [14]

    V2v-posenet: V oxel-to-voxel prediction network for accu- rate 3d hand and human pose estimation from a single depth map

    Gyeongsik Moon, Ju Yong Chang, and Kyoung Mu Lee. V2v-posenet: V oxel-to-voxel prediction network for accu- rate 3d hand and human pose estimation from a single depth map. In Proceedings of the IEEE conference on computer vision and pattern Recognition, pages 5079–5088, 2018. 2

    2018

  7. [15]

    OpenPTrack: People tracking for het- erogeneous networks of color-depth cameras

    Matteo Munaro, Alex Horn, Randy Illum, Jeff Burke, and Radu Bogdan Rusu. OpenPTrack: People tracking for het- erogeneous networks of color-depth cameras. In IAS-13 Workshop Proceedings: 1st Intl. Workshop on 3D Robot Per- ception with Point Cloud Library, pages 235–247. Citeseer,

  8. [16]

    PointV oxel: A Simple and Effective Pipeline for Multi-View Multi-Modal 3D Human Pose Estimation

    Zhiyu Pan, Zhicheng Zhong, Wenxuan Guo, Yifan Chen, Jianjiang Feng, and Jie Zhou. PointV oxel: A Simple and Effective Pipeline for Multi-View Multi-Modal 3D Human Pose Estimation. arXiv preprint arXiv:2312.06409, 2023. 2

    2023 arXiv

  9. [17]

    Tessetrack: End-to- end learnable multi-person articulated 3d pose tracking

    N Dinesh Reddy, Laurent Guigues, Leonid Pishchulin, Jayan Eledath, and Srinivasa G Narasimhan. Tessetrack: End-to- end learnable multi-person articulated 3d pose tracking. In Proceedings of the IEEE/CVF conference on computer vi- sion and pattern recognition, pages 15190–15200...

    2021

  10. [18]

    Multiple Kinect based system to monitor and analyze key performance indi- cators of physical training

    Karolis Ryselis, Tautvydas Petkus, Tomas Bla ˇzauskas, Rytis Maskeli¯unas, and Robertas Dama ˇseviˇcius. Multiple Kinect based system to monitor and analyze key performance indi- cators of physical training. Human-Centric Computing and Information Sciences, 10:1–22, 2020. 2

    2020

  11. [19]

    MVOR: A multi-view RGB-D operating room dataset for 2D and 3D human pose estimation

    Vinkle Srivastav, Thibaut Issenhuth, Abdolrahim Kadkho- damohammadi, Michel de Mathelin, Afshin Gangi, and Nicolas Padoy. MVOR: A multi-view RGB-D operating room dataset for 2D and 3D human pose estimation. arXiv preprint arXiv:1808.08180, 2018. 4

    2018 arXiv

  12. [20]

    SelfPose3d: Self-Supervised Multi-Person Multi-View 3d Pose Estima- tion

    Vinkle Srivastav, Keqi Chen, and Nicolas Padoy. SelfPose3d: Self-Supervised Multi-Person Multi-View 3d Pose Estima- tion. In Proceedings of the IEEE/CVF Conference on Com- puter Vision and Pattern Recognition (CVPR), pages 2502– 2512, 2024. 2

    2024

  13. [21]

    Iterative Greedy Matching for 3D Human Pose Tracking from Multiple Views

    Julian Tanke and Juergen Gall. Iterative Greedy Matching for 3D Human Pose Tracking from Multiple Views. In German Conference on Pattern Recognition, 2019. 2

    2019

  14. [22]

    V oxelPose: Towards Multi-Camera 3D Human Pose Estimation in Wild Environment

    Hanyue Tu, Chunyu Wang, and Wenjun Zeng. V oxelPose: Towards Multi-Camera 3D Human Pose Estimation in Wild Environment. In European Conference on Computer Vision (ECCV), 2020. 2

    2020

  15. [23]

    Direct Multi-view Multi-person 3D Human Pose Estimation

    Tao Wang, Jianfeng Zhang, Yujun Cai, Shuicheng Yan, and Jiashi Feng. Direct Multi-view Multi-person 3D Human Pose Estimation. Advances in Neural Information Process- ing Systems, 2021. 2

    2021

  16. [24]

    Graph-based 3d multi-person pose estimation using multi-view images

    Size Wu, Sheng Jin, Wentao Liu, Lei Bai, Chen Qian, Dong Liu, and Wanli Ouyang. Graph-based 3d multi-person pose estimation using multi-view images. In ICCV, 2021. 2

    2021

  17. [25]

    Faster V oxelPose: Real-time 3D Human Pose Estima- tion by Orthographic Projection

    Hang Ye, Wentao Zhu, Chunyu Wang, Rujie Wu, and Yizhou Wang. Faster V oxelPose: Real-time 3D Human Pose Estima- tion by Orthographic Projection. In European Conference on Computer Vision (ECCV), 2022. 2 6

    2022

Pith tools

Reviewed August 9, 2026 · model on record in the stance chip above.