mlsanityData to 7 Oct 2026

Contrastive Positive Mining for Unsupervised 3D Action Representation Learning

arXiv 2208.03497 ↗First posted 6 Aug 2022cs.CV cs.MMcited by 34

Not on the map. Why?

Abstract (arXiv)

Recent contrastive based 3D action representation learning has made great progress. However, the strict positive/negative constraint is yet to be relaxed and the use of non-self positive is yet to be explored. In this paper, a Contrastive Positive Mining (CPM) framework is proposed for unsupervised skeleton 3D action representation learning. The CPM identifies non-self positives in a contextual queue to boost learning. Specifically, the siamese encoders are adopted and trained to match the similarity distributions of the augmented instances in reference to all instances in the contextual queue. By identifying the non-self positive instances in the queue, a positive-enhanced learning strategy is proposed to leverage the knowledge of mined positives to boost the robustness of the learned latent space against intra-class and inter-class diversity. Experimental results have shown that the proposed CPM is effective and outperforms the existing state-of-the-art unsupervised methods on the challenging NTU and PKU-MMD datasets.

Key facts. Cited by 34 papers in the map (data through 7 Oct 2026). Builds on 31 of its 42 references that are in the map. First posted on arXiv 6 Aug 2022.

Builds on

31 of 42 references.

References · 31

  1. 2020A Simple Framework for Contrastive Learning of Visual Representationscited by 11,979
  2. 2018Representation Learning with Contrastive Predictive Codingcited by 7,662
  3. 2019Momentum Contrast for Unsupervised Visual Representation Learningcited by 7,482
  4. 2016SGDR: Stochastic Gradient Descent with Warm Restartscited by 4,993
  5. 2020Bootstrap your own latent: A new approach to self-supervised Learningcited by 4,712
  6. 2020Exploring Simple Siamese Representation Learningcited by 2,853
  7. 2021Barlow Twins: Self-Supervised Learning via Redundancy Reductioncited by 1,818
  8. 2016Unsupervised Learning of Visual Representations by Solving Jigsaw Puzzlescited by 1,379
  9. 2019Contrastive Multiview Codingcited by 1,297
  10. 2018Spatial Temporal Graph Convolutional Networks for Skeleton-Based Action Recognitioncited by 1,199
  11. 2016NTU RGB+D: A Large Scale Dataset for 3D Human Activity Analysiscited by 683
  12. 2019NTU RGB+D 120: A Large-Scale Benchmark for 3D Human Activity Understandingcited by 530
19 more
  1. 2017Large Batch Training of Convolutional Networkscited by 529
  2. 2017Mean teachers are better role models: Weight-averaged consistency targets improve semi-supervised deep learning resultscited by 470
  3. 2015Unsupervised Learning of Video Representations using LSTMscited by 434
  4. 2019An Attention Enhanced Graph Convolutional LSTM Network for Skeleton-Based Action Recognitioncited by 151
  5. 2021SEED: Self-supervised Distillation For Visual Representationcited by 132
  6. 2017A New Representation of Skeleton Sequences for 3D Action Recognitioncited by 125
  7. 20213D Human Action Representation Learning via Cross-View Consistency Pursuitcited by 105
  8. 2020MS2L: Multi-Task Self-Supervised Learning for Skeleton Based Action Recognitioncited by 92
  9. 2019PREDICT & CLUSTER: Unsupervised Skeleton Based Action Recognitioncited by 90
  10. 2021Skeleton-Contrastive 3D Action Representation Learningcited by 76
  11. 2020Augmented Skeleton Based Contrastive Action Learning with Momentum LSTM for Unsupervised Action Recognitioncited by 59
  12. 2018Iterative Reorganization with Weak Spatial Constraints: Solving Arbitrary Jigsaw Puzzles for Unsupervised Representation Learningcited by 51
  13. 2017Unsupervised Learning of Long-Term Motion Dynamics for Videoscited by 37
  14. 2018Unsupervised Feature Learning of Human Actions as Trajectories in Pose Embedding Manifoldcited by 26
  15. 2019Skeleton Image Representation for 3D Action Recognition based on Tree Structure and Reference Jointscited by 25
  16. 2017Scene Flow to Action Map: A New Representation for RGB-D based Action Recognition with Convolutional Neural Networkscited by 20
  17. 2018Action Recognition for Depth Video using Multi-view Dynamic Imagescited by 18
  18. 2017Learning and Refining of Privileged Information-based RNNs for Action Recognition from Depth Sequencescited by 8
  19. 2020SAR-NAS: Skeleton-based Action Recognition via Neural Architecture Searchingcited by 3

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2022: 1; 2023: 9; 2024: 9; 2025: 10; 2026: 5.

34 papers.

Sort:

By area:

  1. 2023Actionlet-Dependent Contrastive Learning for Unsupervised Skeleton-Based Action Recognitioncited by 41
  2. 2023Masked Motion Predictors are Strong 3D Action Representation Learnerscited by 41
  3. 2022ViA: View-invariant Skeleton Action Representation Learning via Motion Retargetingcited by 22
  4. 2023Prompted Contrast with Masked Motion Modeling: Towards Versatile 3D Action Representation Learningcited by 19
  5. 2023Self-Supervised 3D Action Representation Learning with Skeleton Cloud Colorizationcited by 16
  6. 2024MacDiff: Unified Skeleton Modeling with Masked Conditional Diffusioncited by 15
  7. 2024USDRL: Unified Skeleton-Based Dense Representation Learning with Multi-Grained Feature Decorrelationcited by 15
  8. 2025Heterogeneous Skeleton-Based Action Representation Learningcited by 12
  9. 2023Pyramid Self-attention Polymerization Learning for Semi-supervised Skeleton-based Action Recognitioncited by 11
  10. 2023SCD-Net: Spatiotemporal Clues Disentanglement Network for Self-supervised Skeleton-based Action Recognitioncited by 11
  11. 2024Vision-Language Meets the Skeleton: Progressively Distillation with Cross-Modal Knowledge for 3D Action Representation Learningcited by 11
  12. 2025Foundation Model for Skeleton-Based Human Action Understandingcited by 11
22 more
  1. 2024Identity-free Artificial Emotional Intelligence via Micro-Gesture Understandingcited by 8
  2. 2024Idempotent Unsupervised Representation Learning for Skeleton-Based Action Recognitioncited by 7
  3. 2025SiMHand: Mining Similar Hands for Large-Scale 3D Hand Pose Pre-trainingcited by 7
  4. 2025Skeleton Motion Words for Unsupervised Skeleton-Based Temporal Action Segmentationcited by 6
  5. 2024Self-Supervised Skeleton-Based Action Representation Learning: A Benchmark and Beyondcited by 5
  6. 2025Spatio-Temporal Joint Density Driven Learning for Skeleton-Based Action Recognitioncited by 4
  7. 2024STARS: Self-supervised Tuning for 3D Action Recognition in Skeleton Sequencescited by 3
  8. 2025Towards Efficient General Feature Prediction in Masked Skeleton Modelingcited by 3
  9. 2024Skeleton2vec: A Self-supervised Learning Framework with Contextualized Target Representations for Skeleton Sequencecited by 2
  10. 2024Pre-Training for 3D Hand Pose Estimation with Contrastive Learning on Large-Scale Hand Images in the Wildcited by 2
  11. 2023Attack-Augmentation Mixing-Contrastive Skeletal Representation Learningcited by 1
  12. 2023Elevating Skeleton-Based Action Recognition with Efficient Multi-Modality Self-Supervisioncited by 1
  13. 2023I2MD: 3D Action Representation Learning with Inter- and Intra-modal Mutual Distillationcited by 1
  14. 2025Cross-Modal Consistency Learning for Sign Language Recognitioncited by 1
  15. 2025Multimodal Skeleton-Based Action Representation Learning via Decomposition and Compositioncited by 1
  16. 2025Learning Transformation-Isomorphic Latent Space for Accurate Hand Pose Estimationcited by 0
  17. 2025MS-CLR: Multi-Skeleton Contrastive Learning for Human Action Recognitioncited by 0
  18. 2026Skeleton-to-Image Encoding: Enabling Skeleton Representation Learning via Vision-Pretrained Modelscited by 0
  19. 2026Unsupervised Skeleton-Based Action Segmentation via Hierarchical Spatiotemporal Vector Quantizationcited by 0
  20. 2026Toward Universal Skeleton-Based Action Recognition across Heterogeneous Skeletons and Open Vocabulariescited by 0
  21. 2026One for All: Generalist Foundation Model for Cross-Sensor Skeleton Representation Learningcited by 0
  22. 2026DiMoP: Diffusion-Driven Motion Representation Learning With Frame-Level Pseudo-Classification for Skeleton-Based Action Recognitioncited by 0