Driving is a concern in today’s society because it has contributed significantly to traffic accidents. In this article, we present a system that recognizes distracted drivers using Convolutional Neural Networks (CNNs) and drowsiness detection using YOLO algorithm and sounds an alarm to get their attention back on the road. In order to help identify distracted driver, this technology also shows the driver’s phone number and vehicle information. Our model takes picturesof the driver and their surroundings in real time using a camera that is mounted in the car. We are able to recognize indicators of driving, like texting or using a mobile device, through analysis with CNNs. The customized message promotes safe driving practices, draws attention to the dangers of driving, and stresses the significance of getting back into the habit of driving. Access to a database with records of driving incidents and vehicle information is available to law enforcement agencies. This makes it possible for them to efficiently observe and track driving behaviors. When the system was evaluated using a real-world driving dataset, the algorithm identified instances of driving withan accuracy rate of more than $96 \%$ and for the drowsiness detection the accuracy goes around $94 \%$. It is simple to use and accessible for drivers with the features like phone number display, car information, and WhatsApp connectivity. This model can, in general, lower the number of driving-related collisions, improve traffic safety, and encourage safe driving practices. Its objective is to increase traffic safety and avoid collisions.
Automatic Detection of Driver Drowsiness and Distraction for Public Safety: Deep Learning Based Approach
2024-04-18
564664 byte
Conference paper
Electronic Resource
English
Driver Distraction and Drowsiness Detection System
IEEE | 2021
|Driver Drowsiness Detection for Road Safety Using Deep Learning
Springer Verlag | 2023
|Driver Drowsiness Detection Using Deep Learning
IEEE | 2021
|Driver Drowsiness Detection Using Deep Learning
IEEE | 2023
|