Research map / Video / Action recognition
Light but Sharp: SlimSTAD for Real-Time Action Detection from Sensor Data
Wei Cui, Lukai Fan, Zhenghua Chen, Min Wu, Shili Xiang, Haixia Wang, Bing Li
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 graph-based channel aggregation (ablation) | SlimSTAD DCM encoder · Sensor dataset | mAP@Avg (%): 89.7 -> 92.3 | Improved |
| add refinement stage to the cascade predictor (ablation) | SlimSTAD · Sensor dataset | mAP@Avg (%): 88.8 -> 92.3 | Improved |
| use DCM encoder instead of R3D backbone (ablation) | SlimSTAD · Sensor dataset | mAP@Avg (%): 90.8 -> 92.3 | Improved |
| use DCM encoder instead of I3D (ablation) | SlimSTAD · Sensor dataset | mAP@Avg (%): 90.1 -> 92.3 | Improved |
| use DCM encoder instead of R(2+1)D (ablation) | SlimSTAD · Sensor dataset | mAP@Avg (%): 84.1 -> 92.3 | Improved |
| apply FP16 quantization | DPWiT · Sensor dataset | mAP@Avg (%): 87.5 -> 86.8 | Hurt |
| apply knowledge distillation | DPWiT · Sensor dataset | mAP@Avg (%): 87.5 -> 82.1 | Hurt |
| apply channel pruning | DPWiT · Sensor dataset | mAP@Avg (%): 87.5 -> 75.8 | Hurt |
| use video-based TAD models directly on sensor data | DyFADet/TADTR · WiFi-HAR | mAP@Avg (%): 58.0 vs 79.7 | Hurt |
| replace I3D-based STADe with DCM encoder | SlimSTAD vs STADe · WiFi-HAR | GFLOPs: 158.3 -> 31.8 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- DyFADet — DyFADet: Dynamic Feature Aggregation for Temporal Action Detection (ECCV 2024) · better on 2 of 2 shared results
- TriDet — TriDet: Temporal Action Detection With Relative Boundary Modeling (CVPR 2023) · better on 1 of 1 shared result
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- STADe: Sensory Temporal Action Detection via Temporal-Spectral Representation Learning (TPAMI 2025 · similarity 0.87)
- Text-Infused Attention and Foreground-Aware Modeling for Zero-Shot Temporal Action Detection (NeurIPS 2024 · similarity 0.85)
- HARDVS: Revisiting Human Activity Recognition with Dynamic Vision Sensors (AAAI 2024 · similarity 0.84)
- TIM: A Time Interval Machine for Audio-Visual Action Recognition (CVPR 2024 · similarity 0.83)
- Temporal Action Detection Model Compression by Progressive Block Drop (CVPR 2025 · similarity 0.83)
- Dual DETRs for Multi-Label Temporal Action Detection (CVPR 2024 · similarity 0.83)
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.
