Cross-Camera Distracted Driver Classification through Feature Disentanglement and Contrastive Learning
Summary: arXiv:2411.13181v3 Announce Type: replace-cross
Abstract
The classification of distracted drivers is pivotal for ensuring safe driving. Previous studies demonstrated the effectiveness of neural networks in automatically predicting driver distraction, fatigue, and potential hazards. However, recent research has uncovered a significant loss of accuracy in these models when applied to samples acquired under conditions that differ from the training data.
In this paper, we introduce a robust model designed to withstand changes in camera position within the vehicle. Our Driver Behavior Monitoring Network (DBMNet) relies on a lightweight backbone and integrates a disentanglement module to discard camera view information from features, coupled with contrastive learning to enhance the encoding of various driver actions.
Key Features of DBMNet
- Disentanglement Module: This innovative feature extracts critical information while removing irrelevant camera view data.
- Contrastive Learning: This technique optimizes the representation of different driver actions, facilitating better classification accuracy.
- Lightweight Backbone: Ensures that the model is efficient without compromising performance.
Experimental Validation
Experiments conducted using a leave-one-camera-out protocol on the daytime and nighttime subsets of the 100-Driver dataset validate the effectiveness of our approach. Cross-dataset and cross-camera experiments conducted on three benchmark datasets, namely AUCDD-V1, EZZ2021, and SFD, demonstrate the superior generalization capabilities of the proposed method.
Performance Metrics
Overall, DBMNet achieves an improvement of 7% in Top-1 accuracy compared to existing efficient approaches. The model’s performance was also evaluated in practical deployment scenarios:
- Deployment on Coral Dev Board: A quantized version of DBMNet was deployed, showcasing its efficiency in real-world applications.
- Lowest Average Error: In comparison to alternatives, DBMNet exhibited the lowest average error rate.
- Compact Model Size: The model maintains a small footprint, making it suitable for embedded systems.
- Fast Inference Time: Ensures real-time processing capabilities essential for driver monitoring applications.
- Minimal Power Consumption: Designed to operate efficiently, making it ideal for long-term use in vehicles.
Conclusion
The advancements presented in this paper underscore the potential of DBMNet to significantly improve the classification of distracted drivers across varying camera positions and lighting conditions. The results not only enhance safety in driving environments but also pave the way for future research in driver behavior monitoring technologies.