Research map / Label-efficient and robust learning / Domain adaptation and generalization
Cross-Domain Multi-Person Human Activity Recognition via Near-Field Wi-Fi Sensing
Xin Li, Jingzhi Hu, Yinghui He, Hongbo Wang, Jin Gan, Jun Luo
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 inter-class margin enlarging in PT and FT (ablation) | WiAnchor · NFS-Fi, RT/HS absent | accuracy without FT samples (%): 54.4 -> 86.3 | Improved |
| add inter-class margin enlarging in PT and FT (ablation) | WiAnchor · NFS-Fi, categories with FT samples | accuracy (%): 85.2 -> 91.4 | Improved |
| add anchor matching (ablation) | WiAnchor · NFS-Fi, RT/HS absent | accuracy without FT samples (%): 52.3 -> 86.3 | Improved |
| add anchor matching (ablation) | WiAnchor · NFS-Fi, categories with FT samples | accuracy (%): 86.7 -> 91.4 | Improved |
| add composite inference strategy (ablation) | WiAnchor · NFS-Fi, RT/HS absent | accuracy without FT samples (%): 79.2 -> 86.3 | Improved |
| add composite inference strategy (ablation) | WiAnchor · NFS-Fi, categories with FT samples | accuracy (%): 90.7 -> 91.4 | Improved |
| WiAnchor instead of standard fine-tuning for absent categories | base GRU model · NFS-Fi, RT and HS absent | accuracy (%): 29.5 -> 86.3 (about 56.8% improvement) | Improved |
| WiAnchor instead of class-sensitive learning | class-sensitive baseline · NFS-Fi | overall accuracy (%): 77.7 -> 90.4 | Improved |
| WiAnchor instead of data augmentation | data augmentation baseline · NFS-Fi | overall accuracy (%): 81.6 -> 90.4 | Improved |
| WiAnchor instead of module optimization | module optimization baseline · NFS-Fi | overall accuracy (%): 79.4 -> 90.4 | Improved |
| increase number of source subjects in pre-training beyond six | WiAnchor · NFS-Fi | accuracy | No clear effect |
| increase FT samples per category beyond 10 | WiAnchor · NFS-Fi | accuracy | No clear effect |
Closest papers by meaning
- ConSense: Continually Sensing Human Activity with WiFi via Growing and Picking (AAAI 2025 · similarity 0.90)
- CSI-Bench: A Large-Scale In-the-Wild Dataset for Multi-task WiFi Sensing (NeurIPS 2025 · similarity 0.85)
- WiFi CSI Based Temporal Activity Detection via Dual Pyramid Network (AAAI 2025 · similarity 0.84)
- Decomposing and Fusing Intra- and Inter-Sensor Spatio-Temporal Signal for Multi-Sensor Wearable Human Activity Recognition (AAAI 2025 · similarity 0.83)
- Multitarget Device-Free Localization via Cross-Domain Wi-Fi RSS Training Data and Attentional Prior Fusion (AAAI 2024 · similarity 0.82)
- UniMTS: Unified Pre-training for Motion Time Series (NeurIPS 2024 · similarity 0.81)
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.
