mlsanityData to 7 Oct 2026

Weakly-supervised Pre-training for 3D Human Pose Estimation via Perspective Knowledge

arXiv 2211.11983 ↗First posted 22 Nov 2022cs.CVcited by 6

Not on the map. Why?

Abstract (arXiv)

Modern deep learning-based 3D pose estimation approaches require plenty of 3D pose annotations. However, existing 3D datasets lack diversity, which limits the performance of current methods and their generalization ability. Although existing methods utilize 2D pose annotations to help 3D pose estimation, they mainly focus on extracting 2D structural constraints from 2D poses, ignoring the 3D information hidden in the images. In this paper, we propose a novel method to extract weak 3D information directly from 2D images without 3D pose supervision. Firstly, we utilize 2D pose annotations and perspective prior knowledge to generate the relationship of that keypoint is closer or farther from the camera, called relative depth. We collect a 2D pose dataset (MCPC) and generate relative depth labels. Based on MCPC, we propose a weakly-supervised pre-training (WSP) strategy to distinguish the depth relationship between two points in an image. WSP enables the learning of the relative depth of two keypoints on lots of in-the-wild images, which is more capable of predicting depth and generalization ability for 3D human pose estimation. After fine-tuning on 3D pose datasets, WSP achieves state-of-the-art results on two widely-used benchmarks.

Key facts. Cited by 6 papers in the map (data through 7 Oct 2026). Builds on 46 of its 64 references that are in the map. First posted on arXiv 22 Nov 2022.

Builds on

46 of 64 references.

References · 46

  1. 2015Deep Residual Learning for Image Recognitioncited by 51,809
  2. 2014Microsoft COCO: Common Objects in Contextcited by 17,498
  3. 2019Deep High-Resolution Representation Learning for Human Pose Estimationcited by 1,339
  4. 2016Stacked Hourglass Networks for Human Pose Estimationcited by 913
  5. 2018Simple Baselines for Human Pose Estimation and Trackingcited by 502
  6. 2016Monocular 3D Human Pose Estimation In The Wild Using Improved CNN Supervisioncited by 493
  7. 2020End-to-End Human Pose and Mesh Reconstruction with Transformerscited by 417
  8. 2017A simple yet effective baseline for 3d human pose estimationcited by 401
  9. 20183D human pose estimation in video with temporal convolutions and semi-supervised trainingcited by 401
  10. 2017Learning from Synthetic Humanscited by 358
  11. 2017Cascaded Pyramid Network for Multi-Person Pose Estimationcited by 341
  12. 2017Integral Human Pose Regressioncited by 255
34 more
  1. 2018CrowdPose: Efficient Crowded Scenes Pose Estimation and A New Benchmarkcited by 214
  2. 2019Semantic Graph Convolutional Networks for 3D Human Pose Regressioncited by 204
  3. 2017Single-Shot Multi-Person 3D Pose Estimation From Monocular RGBcited by 195
  4. 2017PoseTrack: A Benchmark for Human Pose Estimation and Trackingcited by 173
  5. 2019Learnable Triangulation of Human Posecited by 160
  6. 2019Camera Distance-aware Top-down Approach for 3D Multi-person Pose Estimation from a Single RGB Imagecited by 131
  7. 2021Human Pose Regression with Residual Log-likelihood Estimationcited by 131
  8. 2017Compositional Human Pose Regressioncited by 124
  9. 2017Towards 3D Human Pose Estimation in the Wild: a Weakly-supervised Approachcited by 114
  10. 2018Ordinal Depth Supervision for 3D Human Pose Estimationcited by 107
  11. 2019Self-Supervised Learning of 3D Human Pose using Multi-view Geometrycited by 98
  12. 2020SRNet: Improving Generalization in 3D Human Pose Estimation with a Split-and-Recombine Approachcited by 94
  13. 2020Detailed 2D-3D Joint Representation for Human-Object Interactioncited by 93
  14. 2019Cross View Fusion for 3D Human Pose Estimationcited by 90
  15. 2018LCR-Net++: Multi-person 2D and 3D Pose Detection in Natural Imagescited by 88
  16. 20183D Human Pose Estimation in the Wild by Adversarial Learningcited by 86
  17. 2019RepNet: Weakly Supervised Training of an Adversarial Reprojection Network for 3D Human Pose Estimationcited by 85
  18. 2021Probabilistic Monocular 3D Human Pose Estimation with Normalizing Flowscited by 84
  19. 20163D Human Pose Estimation from a Single Image via Distance Matrix Regressioncited by 73
  20. 2016Synthesizing Training Images for Boosting Human 3D Pose Estimationcited by 52
  21. 2020SMAP: Single-Shot Multi-Person Absolute 3D Pose Estimationcited by 52
  22. 2019In the Wild Human Pose Estimation Using Explicit 2D Features and Intermediate 3D Representationscited by 48
  23. 2020HMOR: Hierarchical Multi-Person Ordinal Relations for Monocular Multi-Person 3D Pose Estimationcited by 44
  24. 2016MoCap-guided Data Augmentation for 3D Pose Estimation in the Wildcited by 37
  25. 2020Predicting Camera Viewpoint Improves Cross-dataset Generalization for 3D Human Pose Estimationcited by 36
  26. 2017Generating Multiple Diverse Hypotheses for Human 3D Pose Consistent with 2D Joint Detectionscited by 30
  27. 2021Action Transformer: A Self-Attention Model for Short-Time Pose-Based Human Action Recognitioncited by 24
  28. 2019Weakly-Supervised Discovery of Geometry-Aware Representation for 3D Human Pose Estimationcited by 23
  29. 2019Multi-Person 3D Human Pose Estimation from Monocular Imagescited by 19
  30. 2021Low-resolution Human Pose Estimationcited by 10
  31. 2022Dynamic Graph Reasoning for Multi-person 3D Pose Estimationcited by 9
  32. 20213D Pose Estimation and Future Motion Prediction from 2D Imagescited by 8
  33. 2017Multi-modal Face Pose Estimation with Multi-task Manifold Deep Learningcited by 6
  34. 2019An End-to-end Framework for Unconstrained Monocular 3D Hand Pose Estimationcited by 2

Built on by

Cited by year

Citing papers per year: 2023: 3; 2024: 2; 2025: 1.

6 papers.

  1. 2023PSVT: End-to-End Multi-person 3D Pose and Shape Estimation with Progressive Video Transformerscited by 31
  2. 2023Human Pose-based Estimation, Tracking and Action Recognition with Deep Learning: A Surveycited by 6
  3. 2024Pre-training a Density-Aware Pose Transformer for Robust LiDAR-based 3D Human Pose Estimationcited by 4
  4. 2023Bridging the Semantic-Numerical Gap: A Numerical Reasoning Method of Cross-modal Knowledge Graph for Material Property Predictioncited by 1
  5. 2024Comparison of Kinematics and Kinetics Between OpenCap and a Marker-Based Motion Capture System in Cyclingcited by 1
  6. 2025StarPose: 3D Human Pose Estimation via Spatial-Temporal Autoregressive Diffusioncited by 1