The Fatigue Driving Detection Model Based on the Combination of CNN and SE Modules

Fatigue driving is one of the leading causes of traffic accidents, posing a serious threat to road safety. Traditional Convolutional Neural Networks (CNN) models have been widely used for fatigue driving detection due to their excellent feature learning capabilities. However, CNN's detection ability may be affected in the presence of complex background noise or occlusions. To address this issue, this paper proposes a fatigue driving detection model that combines CNN with Squeeze-and-Excitation (SE) modules. While introducing the SE module, the depth of the CNN model is appropriately increased and parameters are tuned, enhancing the model's robustness and applicability in complex environments. The model utilizes CNN's powerful facial feature extraction capabilities to recognize key facial features, while the SE module dynamically adjusts the importance of each feature channel, thereby increasing the focus on critical features. Experimental results show that the proposed SeNet model achieves an accuracy of approximately 98.8% on a small-scale dataset, significantly outperforming the CNN without the SE module (94.8%). This model can effectively detect the driver's fatigue state, providing a practical solution for improving road safety and reducing traffic accidents caused by fatigue driving, with significant social and traffic safety implications.