Design and Implementation of a Fatigue Detection System Based on Dlib for Driver Facial Features.

There are currently three commonly used methods for detecting fatigue driving, namely physiological feature based detection, vehicle driving information based detection, and driver behavior feature based detection. The first two methods have issues such as invasiveness to drivers and low detection accuracy, which makes them unable to be put into large-scale use under actual production conditions. This article adopts a driver fatigue detection method based on Dlib, using the 68 person face feature point model database contained in the Dlib database for facial localization. The driver is judged whether they are tired driving by the three fatigue features of blinking, yawning, and nodding. The three features are standardized, and the number of blinking, nodding, and yawning is counted using coordinate points to determine whether they are tired. The recognition success rate is about 95%, More than 90% reach the application level.