Research map / Large language models / AI and society
EgoEMS: A High-Fidelity Multimodal Egocentric Dataset for Cognitive Assistance in Emergency Medical Services
Keshara Weerasinghe, Xueren Ge, Tessa Heick, Lahiru Nuwan Wijayasingha, Anthony Cortez, Abhishek Satpathy, John Stankovic, Homa Alemzadeh
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 |
|---|---|---|---|
| early fusion of smartwatch IMU with ego video (ablation) | MTRSAP transformer · EgoEMS keystep classification (67 classes) | top-1 accuracy: 62.3 -> 62.2 | No clear effect |
| fuse smartwatch IMU with ego video features (ablation) | transformer · EgoEMS keystep segmentation (5 s windows) | accuracy: +6% (best 61%) | Improved |
| use ResNet50 instead of I3D video features (ablation) | MTRSAP · EgoEMS keystep classification (67 classes) | top-1 accuracy: 60.41 -> 62.27 | Improved |
| fuse video and IMU for CPR feedback (ablation) | CPR quality estimation · EgoEMS CPR quality estimation (compression depth) | RMSE | Improved |
| use smartwatch IMU only instead of fusion for compression rate (ablation) | CPR quality estimation · EgoEMS CPR quality estimation (compression rate) | RMSE | Improved |
| use Gemini-2.5 with EMS-context prompt instead of Whisper/Google ASR (ablation) | speech transcription · Speech recognition on 10 EgoEMS samples | WER: 0.59 -> 0.31 | Improved |
| use semi-automatic DETR+SAM2 labeling instead of manual boxes (ablation) | bounding-box annotation · Bounding-box annotation vs manual (10% of trials) | annotation time (hrs): 66 -> 1 | Improved |
| use audio-only zero-shot pipeline instead of video+IMU (ablation) | keystep segmentation · EgoEMS keystep segmentation (5 s windows) | accuracy (%): 61% -> 38% | Hurt |
Closest papers by meaning
- IndEgo: A Dataset of Industrial Scenarios and Collaborative Work for Egocentric Assistants (NeurIPS 2025 · similarity 0.84)
- EgoExOR: An Ego-Exo-Centric Operating Room Dataset for Surgical Activity Understanding (NeurIPS 2025 · similarity 0.83)
- LifeEval: A Multimodal Benchmark for Assistive AI in Egocentric Daily Life Tasks (CVPR 2026 · similarity 0.82)
- EgoExoLearn: A Dataset for Bridging Asynchronous Ego- and Exo-centric View of Procedural Activities in Real World (CVPR 2024 · similarity 0.82)
- Measuring Cross-Modal Interactions in Multimodal Models (AAAI 2025 · similarity 0.81)
- CPR-Coach: Recognizing Composite Error Actions based on Single-class Training (CVPR 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.
