Research map / Video / Action recognition
OMG-Bench: A New Challenging Benchmark for Skeleton-based Online Micro Hand Gesture Recognition
Haochen Chang, Pengfei Ren, Buyuan Zhang, Da Li, Tianhao Han, Haoyang Zhang, Liang Xie, Hongbo Chen, Erwei Yin
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 frame-level memory bank (ablation) | HMATr · OMG-Bench (cross-subject test) | DR: 86.1% -> 89.2% | Improved |
| Add window-level memory bank (ablation) | HMATr · OMG-Bench (cross-subject test) | DR: 87.0% -> 89.2% | Improved |
| Add position-aware queries (ablation) | HMATr · OMG-Bench (cross-subject test) | DR: 82.4% -> 89.2% | Improved |
| Add position-aware interaction module (ablation) | HMATr · OMG-Bench (cross-subject test) | DR: 82.4% -> 89.2% | Improved |
| Add frame memory interaction (ablation) | HMATr · OMG-Bench (cross-subject test) | DR: 87.5% -> 89.2% | Improved |
| Add query memory interaction (ablation) | HMATr · OMG-Bench (cross-subject test) | DR: 85.3% -> 89.2% | Improved |
| End-to-end query-based detection and classification vs best baseline | HMATr · OMG-Bench (cross-subject test) | DR: 81.6% -> 89.2% | Improved |
| Add CTC loss to HD-GCN | HD-GCN · OMG-Bench (cross-subject test) | DR: 77.8% -> 77.3% | No clear effect |
| Two-stage detect-then-classify (Det) vs sliding window | HD-GCN · OMG-Bench (cross-subject test) | DR: 77.8% -> 66.4% | Hurt |
| Apply HMATr to macro gesture dataset | HMATr · SHREC'22 | DR: 0.92 -> 0.95 | Improved |
Closest papers by meaning
- Doodle to Detect: A Goofy but Powerful Approach to Skeleton-based Hand Gesture Recognition (NeurIPS 2025 · similarity 0.88)
- Skeleton-Based Gesture Recognition Using Several Fully Connected Layers with Path Signature Features and Temporal Transformer Module (AAAI 2019 · similarity 0.87)
- Decoupled Representation Learning for Skeleton-Based Gesture Recognition (CVPR 2020 · similarity 0.86)
- Online Detection and Classification of Dynamic Hand Gestures With Recurrent 3D Convolutional Neural Network (CVPR 2016 · similarity 0.85)
- Diverse 3D Hand Gesture Prediction From Body Dynamics by Bilateral Hand Disentanglement (CVPR 2023 · similarity 0.85)
- emg2pose: A Large and Diverse Benchmark for Surface Electromyographic Hand Pose Estimation (NeurIPS 2024 · 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.
