基于脸部特征和头部姿态的疲劳检测方法

针对疲劳驾驶检测方法存在疲劳特征单一、鲁棒性低和不能因驾驶员不同定制疲劳阈值等问题，提出了一种基于脸部特征和头部姿态的疲劳检测方法。利用HOG(histogram of oriented gradients)特征算子和回归树算法进行人脸检测和人脸关键点定位；通过脸部关键点结合坐标系变换估计头部姿态欧拉角；建立深度残差神经网络模型对眼部疲劳特征进行提取，同时结合眼部、嘴部纵横比和头部姿态欧拉角进行疲劳特征提取；利用眼部、嘴部和头部姿态疲劳特征建立针对不同驾驶员的支持向量机模型对疲劳驾驶进行预警。实验表明：在YawDD和自建疲劳模拟数据集上，该方法均表现出较高的准确率和鲁棒性，在某一疲劳特征检测受阻时依然能进行较好的疲劳预警。

Fatigue Detection Method Based on Facial Features and Head Posture

Aiming at the of the single fatigue characteristics, low robustness and inability to customize fatigue thresholds for different drivers of fatigue detection methods, a method based on facial features and head posture is proposed. In face detection and face key point positioning HOG feature operator and regression tree algorithm are used. In head posture estimation, head posture Euler angle is estimated by combining the face key points with the coordinate system transformation. In fatigue feature extraction, a deep residual neural network model is established to extract the eye fatigue features, which the eye, mouth aspect ratio and head posture Euler angle. The fatigue characteristics of eyes, mouth and head are used to establish the support vector machine models for different drivers to provide the early fatigue driving warning. Experiments show that on YawDD and self-built fatigue simulation data sets, the method shows high accuracy and robustness, and can provide better fatigue warning when a certain fatigue feature detection is blocked.