Research On Driver’s Fatigue Detection Method Based On Mouth Features

Currently,the traffic safety issues have become a worldwide wide range of topics.According to statistics,there are about 40%of traffic safety issues caused by the driver fatigue.Therefore.the researchers on the driver fatigue detection and the prevention of traffic accidents have a very important practical significance.Based on the in-depth analysis of the existing relevant driver fatigue detection technology in depth analysis,on the video,we adopt the SVM(Support vector Machine)technology to monitor the driver fatigue based on the mouth state.First of all,the face is tracked from each frame of the collected video based on the KLT(Kanade-Lucas-Tomassi)algorithm to extract the face features.After that,we detect the mouth using the Viola-Jones object detection technique.The detected mouth features(open and closed mouth features)are obtained by the theory of circular tough transform(CHU).Then,by using all these features(open and closed mouth features),we calculate a set of frame images to find the mouth-open time and yawning frequency.To reduce the high time complexity of video-oriented iterative calculation,the data import process is fulfilled in real-time based on a sliding window.The extracted features are finally combined and presented to the support vector machine classifier(SVM),resulting in either one of two possible driver states.The experimental simulation is conducted to test the above algorithms by using the real data.The experimental results showed the effectiveness of this method.