Car Driver's Behaviors Detections Using Ensemble Model

The logistics industry's growth has created an increased demand for transportation, which has resulted in driver fatigue and distraction, leading to accidents. To address this issue, we have developed a Car Driver's Behavior Detection System using deep learning. The system analyzes video data captured from cameras installed inside the passenger compartment and uses YOLOv5, MediaPipe, and LSTM models for analysis. The system's performance is evaluated using F1 scores. The system identifies in-appropriate behavior and alerts the driver, aiming to reduce driver fatigue and distraction and prevent accidents. However, the system's performance may be affected by varying brightness levels in different environments.