Driver Fatigue Detection Method Based on Human Pose Information Entropy

Figure 1

There are four main steps. First, the OpenPose algorithm is applied to extract skeleton information and detect the key points. Then, four types of fatigue features (F_right, F_s, F_{(x, y)}, and F_left) are extracted using the key points. Next, information entropy theory [20] is employed to calculate the entropy value of fatigue features over a specified period. Finally, an SVM classifier is proposed to obtain the fatigue state.