Research map / Video / Action recognition
Idempotent Unsupervised Representation Learning for Skeleton-Based Action Recognition
Lilang Lin, Lehong Wu, Jiahang Zhang, Jiaying Liu
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add manifold decoupled feature fusion module (adapter) (ablation) | IGM · NTU 60 xsub (linear) | accuracy (%): 83.3 -> 85.1 | Improved |
| add feature idempotency constraint to FFM model (ablation) | IGM · NTU 60 xsub (linear) | accuracy (%): 85.1 -> 85.5 | Improved |
| add distribution idempotency constraint to FFM model (ablation) | IGM · NTU 60 xsub (linear) | accuracy (%): 85.1 -> 86.0 | Improved |
| combine feature and distribution idempotency constraints (ablation) | IGM · NTU 60 xsub (linear) | accuracy (%): 86.0 -> 86.2 | Improved |
| add idempotency losses and feature fusion on KNN evaluation (ablation) | IGM · NTU 60 xsub (KNN) | accuracy (%): 64.7 -> 69.3 | Improved |
| use idempotent diffusion generative model instead of MAE-based generative pretraining (MAMP) | skeleton self-supervised learning · NTU 60 xsub | accuracy (%): 84.9 -> 86.2 | Improved |
| use generative denoising at test time for corrupted skeletons | IGM · NTU-C 60 xsub joint noise (1.0, 0.1) | accuracy (%): 12.7 -> 58.7 vs ActCLR | Improved |
| apply idempotency to improve generation quality in mask prediction | IGM · NTU 60 xsub mask prediction | MPJPE (mm): 87.6 -> 79.2 vs MDM | Improved |
| transfer IGM features pretrained on NTU 60 to PKUMMD II | IGM · PKUMMD II xsub | accuracy (%): 55.9 -> 59.8 vs ActCLR | Improved |
Closest papers by meaning
- Self-Supervised Skeleton-Based Action Representation Learning: A Benchmark and Beyond (IJCV 2026 · similarity 0.90)
- Neuron: Learning Context-Aware Evolving Representations for Zero-Shot Skeleton Action Recognition (CVPR 2025 · similarity 0.89)
- Heterogeneous Skeleton-Based Action Representation Learning (CVPR 2025 · similarity 0.89)
- Revealing Key Details to See Differences: A Novel Prototypical Perspective for Skeleton-based Action Recognition (CVPR 2025 · similarity 0.88)
- Skeleton-based Action Recognition with Non-linear Dependency Modeling and Hilbert-Schmidt Independence Criterion (AAAI 2025 · similarity 0.88)
- MaskCLR: Attention-Guided Contrastive Learning for Robust Action Representation Learning (CVPR 2024 · similarity 0.88)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
