Somnolence Detection System Utilizing Deep Neural Network

Numerous studies indicates the importance of drowsiness detection on the road since it can lessen accidents most especially to the commuters who take long hours of travel before reaching their destinations. Somnolence or drowsiness is a state of strong desire for sleep as a result of many factors which include stress and fatigue due to certain types of medication, sleep disorder, and boredom through repetitive tasks such as driving for long periods. Currently, deep learning is widely used in machines or devices. Thus, deep learning is improved to series studies and trends nowadays. Other studies includes both yawning and head tilting only but the created new system includes the detection of patterns in the facial features (eye closure, nodding/head tilting, and yawning) of a driver using captured images of a camera and forwards it to the SqueezeNet deep neural network that achieves faster development and retraining of models. The device has an alarm that alerts the driver when the pattern exhibited by the driver's facial features is analyzed as drowsy. The system percentage result is 97% overall accuracy.