Driver Distraction Detection — Multi-Model System
Detect driver distraction using your choice of advanced models, Visualise model attention with Grad-CAM and explain predictions with LIME.
Select a model and upload a driver image
Available Models:
- Vision Transformer (ViT): ViT-Base fine-tuned on AUC dataset
- Stacking Meta Learner: LogisticRegression meta-learner on top of ViT predictions
- Blending Ensemble: Weighted average of ViT predictions (ViT weight: 0.32)
Select Model
Classification Categories
| Code | Behaviour |
|---|---|
| c0 | Safe Driving |
| c1 | Texting - Right |
| c2 | Talking on Phone - Right |
| c3 | Texting - Left |
| c4 | Talking on Phone - Left |
| c5 | Operating Radio |
| c6 | Drinking |
| c7 | Reaching Behind |
| c8 | Hair and Makeup |
| c9 | Talking to Passenger |
Dataset: AUC Distracted Driver Dataset v2 — Camera 1 | Device: cpu