JingyibySUTsoftware/Yolov5-deepsort-driverDistracted-driving-behavior-detection
A deep-learning system using YOLOv5 and Dlib facial landmark detection to monitor driver fatigue and distracted behaviors such as phone use, smoking, and drinking.
Velocity · 7d
+0.4
★ / day
Trend
→steady
star history
The system performs fatigue detection by analyzing facial landmarks via Dlib to measure eye closure and yawning frequency, applying the PERCLOS model to evaluate drowsiness. It detects distracted behaviors including phone usage, smoking, and drinking using a YOLOv5 object detection model with DeepSort tracking. A PyQt5 interface provides real-time monitoring and visual warnings to the driver.