Drowsy Alert: A System to Detect and Alert Driver's Drowsiness for Road Safety

Driver drowsiness is a leading cause of road accidents, often resulting in severe injuries and fatalities. To mitigate this risk and enhance road safety, we propose Drowsy Alert, an intelligent system designed to detect and alert drivers' drowsiness in real time using a combination of cascade classifiers and convolutional neural networks (CNN). DrowsyAlert leverages a cascade classifier to identify and track the driver's face in video frames captured by an onboard camera. The cascade classifier accurately detects facial landmarks, enabling the system to extract relevant features, including eye closure, head position, and facial expressions. These features are then fed into a CNN model, which is trained to classify drowsiness based on learned patterns from a diverse dataset of drowsy and alert driver images. The system continuously analyzes the driver's facial features, monitoring for signs of drowsiness. If drowsiness is detected, DrowsyAlert promptly generates an alert, notifying the driver through visual, auditory, or haptic feedback. To evaluate the performance of DrowsyAlert, extensive experiments were conducted on a diverse dataset of driver images, encompassing various lighting conditions, driver poses, and levels of drowsiness. The results demonstrate the system's high accuracy in detecting drowsiness and generating timely alerts, thus contributing to improved road safety.