Automatic drowsiness detection for preventing road accidents via 3dgan and three-level attention

Driver drowsiness is one of the reasons for large number of road accidents these days. With the advancement in computer vision technologies, smart/intelligent cameras are developed to identify drowsiness in drivers, thereby alerting drivers which in turn reduce accidents when they are in fatigue. Deep learning is a powerful technique for detecting drowsiness in drivers to prevent road accidents. It uses advanced neural networks to analyze data to detect subtle changes in a driver's facial expressions, eye movements, and head position that may indicate drowsiness. Some of the limitations reviewed from the existing works are false alarm and limited applicability. Hence to overcome these issues, this work proposed Dependent Generative Adversarial Network (DGAN) for detecting drowsiness. With the help of eye blinking, eye closure and also yawning indications, the proposed model will detect whether the person is driving the vehicle in drunken drowsy state. Additionally, the proposed method detect whether the driver is wearing glasses or not and also the lighting conditions. This phase also causes a significant increase in eyes and mouth detection percentage in the next stage. This system alerts driver with an alarm when the driver is in sleepy mood. In order to train the proposed network a custom dataset of about 6000 images was compiled and labeled with the objects face, eye open and eye closed. Out of these, around 1000 images were randomly separated and used to test the trained model. The proposed model can used for various background and environmental changes like indoor, outdoor, day and night. Experimental results confirm that the proposed method efficiently detects the driver behavior. The proposed model achieves a high driver drowsiness with RD of 91.3%, RFA of 7.62% and RA of 91.82%.