Multi-Task Cascaded Convolutional Neural Networks (MTCNN) based Driver Drowsiness Detection

Road accidents are most commonly caused by tired and fatigued drivers. They increase the number of fatalities and injuries globally each year. The proposed method uses three network layers of Multi-Task Cascaded Convolutional Neural Networks (MTCNN) to detect faces. To prevent traffic accidents while driving, it is possible to monitor the driver's level of weariness by observing how frequently his eyes blink. An additional alarm is sounded to signal the automated sidelining and breaking of the vehicle if the driver continues to feel sleepy after the siren. Once the picture pyramid design is supplied to the three network layers, the eye ratio will be calculated by using visual data and Open CV.