Maixdock Based Driver Drowsiness Detection System Using CNN

This article demonstrates how to use Maixdock to build a drowsy driving monitoring system. A behavioral deterioration in one’s ability to drive is known as drowsy driving. To categorize sleepiness signs like breathing and squinting deep learning has been used in this study. Yolo design training was done using example pictures. To categorize sleepiness signs like blinking and breathing, this study employs the Convolutional Neural Network (CNN). The CNN design was trained using 1310 images in total. Then the yolo was trained with Adam’s optimization method. Ten people participated in a live experiment to determine how well this version worked. In this study, a new deep learning-based method for real-time sleepiness monitoring is proposed. It can be easily applied on a low-cost integrated chip and has a good level of performance. A single computer can then receive the data that was gathered. Facial characteristics, such as gaping, and ocular metrics, such as eye-closing, are the areas of concern used here. In this study, additional variables like camera distance from the vehicle and illumination effects are examined. These variables have the potential to influence the rate of categorization accuracy. The key of the car will not turn on if the motorist is intoxicated until the situation is altered. If the vehicle is already in a drivable state, the system will warn the driver via an alarm, and a heartbeat monitor will also identify the data and warn the driver. The Proposed method gives a good accuracy of detection, approximately 90% higher than existing methods.