Research map / Video / Action recognition
DART: Deformable Adaptive Reasoning with Temporal Queries for Online Skeleton-Based Action Recognition
Chaoyang Zheng, Yang Xiao, Tingbing Yan, Jinfang Gan, Xintao Zhang, Ran Wang, Zhiguo Cao, Joey Tianyi Zhou
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 deformable temporal sampling (DMTN) (ablation) | online skeleton baseline · NTU RGB+D 60 X-Sub (online) | AUC (%): 69.25 -> 73.37 | Improved |
| add Temporal Query Decoder on top of DMTN (ablation) | DART · NTU RGB+D 60 X-Sub (online) | AUC (%): 73.37 -> 73.99 | Improved |
| add TQD alone with uniform memory (ablation) | online skeleton baseline · NTU RGB+D 60 X-Sub (online) | AUC (%): 69.25 -> 70.27 | Improved |
| learnable offsets instead of fixed dilation (ablation) | DMTN · NTU RGB+D 60 X-Sub (online) | AUC (%): 72.37 -> 73.37 | Improved |
| increase DMTN layers from 4 to 5 (ablation) | DART · NTU RGB+D 60 X-Sub (online) | AUC (%): 73.37 -> 73.03 | Hurt |
| increase number of queries from 10 to 14 (ablation) | DART · NTU RGB+D 60 X-Sub (online) | AUC (%): 73.99 -> 73.29 | Hurt |
| increase decoder layers from 4 to 5 (ablation) | DART · NTU RGB+D 60 X-Sub (online) | AUC (%): 73.99 -> 73.57 | Hurt |
| replace Neural ODE future prediction with query decoder and deformable offsets | InfoGCN++ · NTU RGB+D 60 X-Sub (online) | FLOPs (G): 10.00G -> 4.59G | Improved |
| DART vs InfoGCN++ on small dataset | DART · NW-UCLA (online) | AUC (%): 85.47 -> 86.34 | Improved |
| DART vs STST at full observation | DART · NTU RGB+D 60 X-Sub (online) | accuracy at 100% observation (%): 91.90 -> 87.63 | Hurt |
Closest papers by meaning
- Neuron: Learning Context-Aware Evolving Representations for Zero-Shot Skeleton Action Recognition (CVPR 2025 · similarity 0.89)
- InfoGCN++: Learning Representation by Predicting the Future for Online Skeleton-Based Action Recognition (TPAMI 2024 · similarity 0.88)
- Boosting Skeleton-based Zero-Shot Action Recognition with Training-Free Test-Time Adaptation (NeurIPS 2025 · similarity 0.87)
- Recovering Complete Actions for Cross-dataset Skeleton Action Recognition (NeurIPS 2024 · similarity 0.87)
- Universal Skeleton Understanding via Differentiable Rendering and MLLMs (ICML 2026 · similarity 0.87)
- Revealing Key Details to See Differences: A Novel Prototypical Perspective for Skeleton-based Action Recognition (CVPR 2025 · similarity 0.86)
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.
