Multi-Fatigue Indicator Detection for Drivers Using FaceMesh

Driver fatigue is a major contributor to road accidents, making real-time and accurate detection essential for transportation safety. This paper proposes a lightweight and stable multi-indicator fatigue detection algorithm based on MediaPipe FaceMesh technology. By extracting and analyzing 3D facial landmarks, the method calculates two key metrics: Eye Aspect Ratio (EAR) for eye closure detection and Mouth Opening Ratio (MOR) for yawning detection. These indicators are fused with temporal thresholds, including continuous eye closure duration and yawning frequency, to improve detection reliability. Experimental validation using the YawDD dataset demonstrates that the proposed dual-feature, dual-index method achieves high accuracy and stability, significantly outperforming traditional rule-based, SVM+HOG, and CNN+LSTM approaches in both false positive and false negative rates. The results verify the method’s feasibility and effectiveness in real-world scenarios.