Key points are not available for this paper at this time.
Drowsiness or sleepiness is a transition stage between alertness and sleep. Drowsy driving causes collisions, injuries and deaths. This paper presents an integrated system for drivers’ drowsiness detection based on deep learning frameworks. The integrated system consists of three parts, i.e., eye region detection, eye state detection and classification, and an alert system generation based on drowsiness level. Faster region based convolutional neural network (f-RCNN) has been employed for eye region detection to extract eye regions from facial images of driver with complicated background. In the next stage, image containing only eye region is fed into a CNN for necessary identification. Finally, an alert system has been generated based on the drowsiness level of drivers’ eye states. The proposed module has been implemented using an Atmega328p microcontroller to achieve a real time evaluation.
Ganguly et al. (Sat,) studied this question.