Research map / Detection and segmentation / 3D detection and driving
MUSE: Multimodal Uncertainty-Based Self-Driven Evolution for Robust Physiological-Signal–Based Driver Fatigue Detection
Jiaheng Wang, Yuan Si, Ang Li, Zhenyu Wang, Tianheng Xu, Honglin Hu
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 |
|---|---|---|---|
| fuse EEG and EOG with MUSE instead of EEG alone (ablation) | MUSE · SEED-VIG fusion comparison (intra-subject) | RMSE: 0.087 -> 0.060 | Improved |
| fuse EEG and EOG with MUSE instead of EOG alone (ablation) | MUSE · SEED-VIG fusion comparison (intra-subject) | RMSE: 0.076 -> 0.060 | Improved |
| fuse EEG and EOG with MUSE instead of EEG alone (ablation) | MUSE · SEED-VIG fusion comparison (inter-subject) | CORR: 0.810 -> 0.866 | Improved |
| use MUSE instead of static concatenation fusion | fusion backbone · SEED-VIG fusion comparison (inter-subject) | RMSE: 0.129 -> 0.097 | Improved |
| use MUSE instead of QMF dynamic fusion | fusion backbone · SEED-VIG fusion comparison (intra-subject) | RMSE: 0.063 -> 0.060 | Improved |
| use MUSE instead of QMF dynamic fusion | fusion backbone · SEED-VIG fusion comparison (inter-subject) | CORR: 0.855 -> 0.866 | Improved |
| dynamic fusion instead of static fusion | fusion methods · SEED-VIG fusion comparison (intra-subject) | RMSE: 0.078-0.079 -> 0.063-0.070 | Improved |
| add KL-divergence regularizer to weight update (ablation) | MUSE · SEED-VIG (MC Dropout evolution) | RMSE and CORR | Improved |
| heteroscedastic regression uncertainty instead of MC Dropout | MUSE · SEED-VIG | latency: 90.31 ms -> 1.62 ms | Improved |
| heteroscedastic regression uncertainty instead of MC Dropout | MUSE · SEED-VIG | FLOPs: 160.95 M -> 2.87 M | Improved |
| swap uncertainty estimator (MC Dropout, deterministic, ensemble, perturbation) | MUSE · SEED-VIG | RMSE and CORR | No clear effect |
| MUSE over best baseline SM model on inter-subject classification | MUSE · SEED-VIG (inter-subject classification) | accuracy (%): 89.8 -> 92.2 | Improved |
Closest papers by meaning
- PhysDrive: A Multimodal Remote Physiological Measurement Dataset for In-vehicle Driver Monitoring (NeurIPS 2025 · similarity 0.87)
- MDFG: Multi-Dimensional Fine-Grained Modeling for Fatigue Detection (AAAI 2025 · similarity 0.85)
- MMTL-UniAD: A Unified Framework for Multimodal and Multi-Task Learning in Assistive Driving Perception (CVPR 2025 · similarity 0.83)
- SafeAuto: Knowledge-Enhanced Safe Autonomous Driving with Multimodal Foundation Models (ICML 2025 · similarity 0.82)
- M3Net: Multimodal Multi-task Learning for 3D Detection, Segmentation, and Occupancy Prediction in Autonomous Driving (AAAI 2025 · similarity 0.82)
- Layer-Wise Modality Decomposition for Interpretable Multimodal Sensor Fusion (NeurIPS 2025 · 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.
