mlsanityData to 7 Oct 2026

Learning Skeletal Graph Neural Networks for Hard 3D Pose Estimation

arXiv 2108.07181 ↗First posted 16 Aug 2021cs.CVcited by 62

Not on the map. Why?

Abstract (arXiv)

Various deep learning techniques have been proposed to solve the single-view 2D-to-3D pose estimation problem. While the average prediction accuracy has been improved significantly over the years, the performance on hard poses with depth ambiguity, self-occlusion, and complex or rare poses is still far from satisfactory. In this work, we target these hard poses and present a novel skeletal GNN learning solution. To be specific, we propose a hop-aware hierarchical channel-squeezing fusion layer to effectively extract relevant information from neighboring nodes while suppressing undesired noises in GNN learning. In addition, we propose a temporal-aware dynamic graph construction procedure that is robust and effective for 3D pose estimation. Experimental results on the Human3.6M dataset show that our solution achieves 10.3% average prediction accuracy improvement and greatly improves on hard poses over state-of-the-art techniques. We further apply the proposed technique on the skeleton-based action recognition task and also achieve state-of-the-art performance. Our code is available at https://github.com/ailingzengzzz/Skeletal-GNN.

Key facts. Cited by 62 papers in the map (data through 7 Oct 2026). Builds on 36 of its 50 references that are in the map. First posted on arXiv 16 Aug 2021.

Builds on

36 of 50 references.

References · 36

  1. 2014Adam: A Method for Stochastic Optimizationcited by 36,960
  2. 2016Semi-Supervised Classification with Graph Convolutional Networkscited by 11,282
  3. 2019Deep High-Resolution Representation Learning for Human Pose Estimationcited by 1,339
  4. 2018Spatial Temporal Graph Convolutional Networks for Skeleton-Based Action Recognitioncited by 1,199
  5. 2016Stacked Hourglass Networks for Human Pose Estimationcited by 913
  6. 2018Representation Learning on Graphs with Jumping Knowledge Networkscited by 875
  7. 2016NTU RGB+D: A Large Scale Dataset for 3D Human Activity Analysiscited by 683
  8. 2019NTU RGB+D 120: A Large-Scale Benchmark for 3D Human Activity Understandingcited by 530
  9. 2019DeepGCNs: Can GCNs Go as Deep as CNNs?cited by 527
  10. 2016Monocular 3D Human Pose Estimation In The Wild Using Improved CNN Supervisioncited by 493
  11. 2019MixHop: Higher-Order Graph Convolutional Architectures via Sparsified Neighborhood Mixingcited by 442
  12. 2017A simple yet effective baseline for 3d human pose estimationcited by 401
24 more
  1. 20183D human pose estimation in video with temporal convolutions and semi-supervised trainingcited by 401
  2. 2017Cascaded Pyramid Network for Multi-Person Pose Estimationcited by 341
  3. 2020Disentangling and Unifying Graph Convolutions for Skeleton-Based Action Recognitioncited by 339
  4. 2019Actional-Structural Graph Convolutional Networks for Skeleton-based Action Recognitioncited by 228
  5. 2019Semantic Graph Convolutional Networks for 3D Human Pose Regressioncited by 204
  6. 2019Semantics-Guided Neural Networks for Efficient Skeleton-Based Human Action Recognitioncited by 135
  7. 2017VNect: Real-time 3D Human Pose Estimation with a Single RGB Cameracited by 129
  8. 20182D/3D Pose Estimation and Action Recognition using Multitask Deep Learningcited by 103
  9. 2019Skeleton-Based Action Recognition with Multi-Stream Adaptive Graph Convolutional Networkscited by 103
  10. 2020SRNet: Improving Generalization in 3D Human Pose Estimation with a Split-and-Recombine Approachcited by 94
  11. 2017Exploiting temporal information for 3D pose estimationcited by 89
  12. 2020Dynamic GCN: Context-enriched Topology Learning for Skeleton-based Action Recognitioncited by 82
  13. 2020Spatial Temporal Transformer Network for Skeleton-based Action Recognitioncited by 76
  14. 2019Learning Graph Convolutional Network for Skeleton-based Human Action Recognition by Neural Searchingcited by 70
  15. 2017Not All Pixels Are Equal: Difficulty-aware Semantic Segmentation via Deep Layer Cascadecited by 53
  16. 2019Trajectory Space Factorization for Deep Video-Based 3D Human Pose Estimationcited by 30
  17. 2018OriNet: A Fully Convolutional Network for 3D Human Pose Estimationcited by 28
  18. 2020Mix Dimension in Poincar'e Geometry for 3D Skeleton-based Action Recognitioncited by 14
  19. 20183D Human Pose Estimation with Relational Networkscited by 10
  20. 2019Generalizing Monocular 3D Human Pose Estimation in the Wildcited by 6
  21. 2020Hop-Hop Relation-aware Graph Neural Networkscited by 5
  22. 2020LookHops: light multi-order convolution and pooling for graph classificationcited by 5
  23. 2019A Unified Deep Framework for Joint 3D Pose Estimation and Action Recognition from a Single RGB Cameracited by 4
  24. 2019Rethinking the Number of Channels for the Convolutional Neural Networkcited by 1

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2021: 4; 2022: 14; 2023: 21; 2024: 13; 2025: 7; 2026: 3.

62 papers.

Sort:

By area:

  1. 2022MixSTE: Seq2seq Mixed Spatio-Temporal Encoder for 3D Human Pose Estimation in Videocited by 158
  2. 2021MHFormer: Multi-Hypothesis Transformer for 3D Human Pose Estimationcited by 151
  3. 2022HuMMan: Multi-Modal 4D Human Dataset for Versatile Sensing and Modelingcited by 115
  4. 2023HumanSD: A Native Skeleton-Guided Diffusion Model for Human Image Generationcited by 104
  5. 2022P-STMO: Pre-Trained Spatial Temporal Many-to-One Model for 3D Human Pose Estimationcited by 80
  6. 2021SmoothNet: A Plug-and-Play Network for Refining Human Poses in Videoscited by 64
  7. 2023Learning Discriminative Representations for Skeleton Based Action Recognitioncited by 62
  8. 2022GFPose: Learning 3D Human Pose Prior with Gradient Fieldscited by 57
  9. 2023SynBody: Synthetic Dataset with Layered Human Models for 3D Human Perception and Modelingcited by 54
  10. 2023GLA-GCN: Global-local Adaptive Graph Convolutional Network for 3D Human Pose Estimation from Monocular Videocited by 46
  11. 2024SkateFormer: Skeletal-Temporal Transformer for Human Action Recognitioncited by 46
  12. 2022STAR-Transformer: A Spatio-temporal Cross Attention Transformer for Human Action Recognitioncited by 43
38 more
  1. 2022DeciWatch: A Simple Baseline for 10x Efficient 2D and 3D Pose Estimationcited by 29
  2. 2024PhysPT: Physics-aware Pretrained Transformer for Estimating Human Dynamics from Monocular Videoscited by 27
  3. 2024MotionCraft: Crafting Whole-Body Motion with Plug-and-Play Multimodal Controlscited by 24
  4. 2023HDFormer: High-order Directed Transformer for 3D Human Pose Estimationcited by 23
  5. 20222D Human Pose Estimation: A Surveycited by 21
  6. 2022GraphMLP: A Graph MLP-Like Architecture for 3D Human Pose Estimationcited by 21
  7. 2022Cross-Modal Learning with 3D Deformable Attention for Action Recognitioncited by 20
  8. 2021Adaptive Multi-view and Temporal Fusing Transformer for 3D Human Pose Estimationcited by 19
  9. 2023Hourglass Tokenizer for Efficient Transformer-Based 3D Human Pose Estimationcited by 15
  10. 2022Skeleton-based Action Recognition via Temporal-Channel Aggregationcited by 13
  11. 2022Leveraging Spatio-Temporal Dependency for Skeleton-Based Action Recognitioncited by 13
  12. 2024Attention-Propagation Network for Egocentric Heatmap to 3D Pose Liftingcited by 13
  13. 2021Hierarchical Graph Networks for 3D Human Pose Estimationcited by 11
  14. 2023Regular Splitting Graph Network for 3D Human Pose Estimationcited by 11
  15. 2023A Single 2D Pose with Context is Worth Hundreds for 3D Human Pose Estimationcited by 11
  16. 2023HSTFormer: Hierarchical Spatial-Temporal Transformers for 3D Human Pose Estimationcited by 9
  17. 2023ActionPrompt: Action-Guided 3D Human Pose Estimation With Text and Pose Promptingcited by 8
  18. 2023HopFIR: Hop-wise GraphFormer with Intragroup Joint Refinement for 3D Human Pose Estimationcited by 7
  19. 2023Lightweight Super-Resolution Head for Human Pose Estimationcited by 7
  20. 2022NIMBLE: A Non-rigid Hand Model with Bones and Musclescited by 6
  21. 2023HTNet: Human Topology Aware Network for 3D Human Pose Estimationcited by 6
  22. 2024GTPT: Group-based Token Pruning Transformer for Efficient Human Pose Estimationcited by 6
  23. 2023TSGCNeXt: Dynamic-Static Multi-Graph Convolution for Efficient Skeleton-Based Action Recognition with Long-term Learning Potentialcited by 5
  24. 2022Towards To-a-T Spatio-Temporal Focus for Skeleton-Based Action Recognitioncited by 4
  25. 2023GATOR: Graph-Aware Transformer with Motion-Disentangled Regression for Human Mesh Recovery from a 2D Posecited by 4
  26. 2023Improving the Robustness of 3D Human Pose Estimation: A Benchmark and Learning from Noisy Inputcited by 4
  27. 2026FMPose3D: monocular 3D pose estimation via flow matchingcited by 3
  28. 20233D Human Pose Lifting with Grid Convolutioncited by 2
  29. 2024Uncertainty-Aware Testing-Time Optimization for 3D Human Pose Estimationcited by 2
  30. 2024SkelFormer: Markerless 3D Pose and Shape Estimation using Skeletal Transformerscited by 2
  31. 2024STGFormer: Spatio-Temporal GraphFormer for 3D Human Pose Estimation in Videocited by 2
  32. 2025NanoHTNet: Nano Human Topology Network for Efficient 3D Human Pose Estimationcited by 2
  33. 2023Iterative Graph Filtering Network for 3D Human Pose Estimationcited by 1
  34. 2023Spatio-temporal MLP-graph network for 3D human pose estimationcited by 1
  35. 2023SoloPose: One-Shot Kinematic 3D Human Pose Estimation with Video Data Augmentationcited by 1
  36. 2024Full-Body Motion Reconstruction with Sparse Sensing from Graph Perspectivecited by 1
  37. 2024Multi-hop graph transformer network for 3D human pose estimationcited by 1
  38. 2024Graph and Skipped Transformer: Exploiting Spatial and Temporal Modeling Capacities for Efficient 3D Human Pose Estimationcited by 1

50 most cited of 62 All 62 (JSON)