A Convolutional Neural Network – Based Fatigue Alert System

Drowsiness detection is a critical application in various fields, particularly in transportation, where driver fatigue is a major cause of accidents. This project aims to develop a simple yet effective system for detecting drowsiness using machine learning techniques. The system leverages computer vision and machine learning algorithms to analyse facial landmarks and eye closure patterns in real-time. We utilize a pre-trained facial feature extraction, focusing on key indicators such as eye aspect ratio (EAR) and head position. The data collected is processed and classified to distinguish between alert and drowsy states. The model is trained on a dataset of annotated facial images, which include both drowsy and non-drowsy conditions. The system's performance is evaluated through accuracy, precision, and recall metrics, demonstrating its potential as a low-cost solution for enhancing safety in vehicles and other settings.