Research map / 3D vision / Depth and stereo
MVDoppler: Unleashing the Power of Multi-View Doppler for MicroMotion-based Gait Classification
Soheil Hor, Shubo Yang, Jae-Ho Choi, Amin Arbabian
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 |
|---|---|---|---|
| use multi-level transformer fusion of two radars instead of single Radar 0 (ablation) | MobileNetV2 · Hand task | accuracy (%): 86.0 -> 90.2 | Improved |
| use late-level transformer fusion of two radars instead of single Radar 0 (ablation) | MobileNetV2 · Distract task | accuracy (%): 63.5 -> 67.4 | Improved |
| use multi-level transformer fusion instead of single Radar 0 (ablation) | MobileNetV2 · Distract task | accuracy (%): 63.5 -> 66.9 | Improved |
| input-level (channel concat) fusion instead of single radar (ablation) | MobileNetV2 · Hand task | accuracy (%): 86.0 -> 85.2 | No clear effect |
| input-level fusion instead of single radar (ablation) | MobileNetV2 · Distract task | accuracy (%): 63.5 -> 60.0 | Hurt |
| multi-level instead of late-level average fusion (ablation) | MobileNetV2 · Hand task | accuracy (%): 87.7 -> 88.5 | Improved |
| multi-level instead of late-level average fusion (ablation) | MobileNetV2 · Distract task | accuracy (%): 65.8 -> 66.9 | Improved |
| transformer instead of average fusion at late level (ablation) | MobileNetV2 · Hand task | accuracy (%): 87.7 -> 88.6 | Improved |
| transformer instead of average fusion at multi level (ablation) | MobileNetV2 · Hand task | accuracy (%): 88.5 -> 90.2 | Improved |
| transformer instead of average fusion at multi level (ablation) | MobileNetV2 · Distract task | accuracy (%): 66.9 -> 66.9 | No clear effect |
| transformer instead of average fusion at late level (ablation) | MobileNetV2 · Distract task | accuracy (%): 65.8 -> 67.4 | Improved |
| longer inference window than a single walking step (ablation) | MobileNetV2 · Hand task | accuracy (%): 85.25% -> 97.9% | Improved |
Closest papers by meaning
- DopplerWild: A Doppler Dataset and Benchmark for Human Kinematic Understanding in the Wild (NeurIPS 2026 · similarity 0.90)
- MVDoppler-Pose: Multi-Modal Multi-View mmWave Sensing for Long-Distance Self-Occluded Human Walking Pose Estimation (CVPR 2025 · similarity 0.84)
- Instantaneous Perception of Moving Objects in 3D (CVPR 2024 · similarity 0.81)
- Exploring More from Multiple Gait Modalities for Human Identification (AAAI 2025 · similarity 0.81)
- HuPerFlow: A Comprehensive Benchmark for Human vs. Machine Motion Estimation Comparison (CVPR 2025 · similarity 0.80)
- Flow Snapshot Neurons in Action: Deep Neural Networks Generalize to Biological Motion Perception (NeurIPS 2024 · similarity 0.80)
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.
