Research map / Video / Action recognition
Recognizing Actions from Robotic View for Natural Human-Robot Interaction
Ziyi Wang, Peiming Li, Hong Liu, Zhichao Deng, Can Wang, Jun Liu, Junsong Yuan, Mengyuan 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 Layered Recognizers (kinematic and action branches) (ablation) | ACTIVE-PC · ACTIVE point cloud | accuracy: 46.76 -> 51.98 | Improved |
| add Multilevel Neighborhood Sampling (ablation) | ACTIVE-PC with Layered Recognizers · ACTIVE point cloud | accuracy: 51.98 -> 55.11 | Improved |
| add Elastic Ellipse Query (ablation) | ACTIVE-PC with LR and MNS · ACTIVE point cloud | accuracy: 55.11 -> 60.10 | Improved |
| add Elastic Ellipse Query to reduce sensitivity to spatial radius (ablation) | ACTIVE-PC · ACTIVE point cloud | accuracy variance across radii: 0.456 -> 0.100 | Improved |
| larger spatial radius r (1.5 to 2) for fixed ball query (ablation) | ACTIVE-PC w/o EEQ · ACTIVE point cloud | accuracy: 55.39 (r=0.5) -> 53.51 (r=2) | Hurt |
| ACTIVE-PC on general NTU RGB+D | ACTIVE-PC · NTU RGB+D cross-subject | accuracy: 91.0 -> 91.7 | Improved |
| add Layered Recognizers to ViT for RGB | ACTIVE-RGB · ACTIVE RGB 8x224 | accuracy: 52.52 -> 55.90 | Improved |
| increase point-cloud input from 8x512 to 12x768 frames x points (ablation) | ACTIVE-PC · ACTIVE point cloud | accuracy: 58.71 -> 60.10 | Improved |
Closest papers by meaning
- RobAVA: A Large-scale Dataset and Baseline Towards Video based Robotic Arm Action Understanding (ICCV 2025 · similarity 0.87)
- JRDB-Act: A Large-Scale Dataset for Spatio-Temporal Action, Social Group and Activity Detection (CVPR 2022 · similarity 0.86)
- ActiveVLA: Injecting Active Perception into Vision-Language-Action Models for Precise 3D Robotic Manipulation (CVPR 2026 · similarity 0.86)
- Human2Robot: Learning Robot Actions from Paired Human-Robot Videos (AAAI 2026 · similarity 0.86)
- Evidential Active Recognition: Intelligent and Prudent Open-World Embodied Perception (CVPR 2024 · similarity 0.85)
- Robots Pre-train Robots: Manipulation-Centric Robotic Representation from Large-Scale Robot Datasets (ICLR 2025 · similarity 0.85)
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.
