Based on Facial Geometric Features and Hand Motion Characteristics Driver Fatigue Detection

Abstract

Fatigue driving is one of the major causes of traffic accidents. In order to solve the problem, a new method based on machine vision for driver fatigue detection is proposed. According to the characteristics of the driver's video image, the regions of interest (ROIs) of face and hand are extracted by the Adaboost algorithm based on skin color detection. Based on SIFT (scale invariant feature transform) feature points matching, the SIFT feature points of the eye, mouth and hand are extracted, and the facial and hand feature parameters are obtained. The 4 characteristic parameters of Perclos, MClosed, Phdown and SA are used as model inputs, and the fatigue grade is used as model output. Three layer BP neural network model is established. The Bayesian regularization and the momentum gradient descent method are used to solve the overfitting phenomena of the traditional BP neural network training with high accuracy and low prediction accuracy. The experimental data show that the method can overcome the influence of illumination, background, angle and individual difference, and the correct recognition rate of fatigue detection is 99.64%.