Ensemble model for Driver Distraction Detection

Accidents on the road are frequently caused by distracted drivers, leading to severe injuries and fatalities. Real-time driver attention detection and mitigation can reduce accidents and boost traffic safety. In this paper, we suggest an ensemble approach for detecting driver distraction using the State Farm Driver Distraction dataset. To accomplish reliable and accurate classification, we combine NasNet-A large, ResNext-101, and EfficientNet-B0 Convolutional Neural Network (CNN) models. Various driver distraction scenarios are shown in the State Farm Driver Distraction dataset, which is a sizable collection of labeled images. The ensemble model aims to achieve higher accuracy and robustness in detecting driver distraction. The ensemble model uses fusion techniques that enhance the overall performance and reliability of the system. We evaluate the performance of our ensemble model from accuracy and loss function. Findings show that the ensemble model performs better than individual CNN models. This research contributes to enhancing road safety by providing an effective real-time system for driver distraction detection, which can be integrated into vehicles and driver assistance systems to mitigate the risks associated with distracted driving.