Paper Key : IRJ************695
Author: Sasirekha.k,Sooriya Narayanan A J,Nishanth M,Mohin P S
Date Published: 03 Nov 2024
Abstract
Driver drowsiness detection is crucial for preventing accidents caused by fatigue. This project presents a non-intrusive detection system using transfer learning with VGG-16 for feature extraction and Light GBM for classification. The model achieves 99% accuracy and processes images in 0.00829 seconds, ensuring real-time detection. It monitors eye movement behavior to identify drowsiness, offering a more comfortable alternative to traditional sensor-based methods. The system performs well under various conditions, enhancing road safety. Future improvements will focus on incorporating additional driver behaviors and optimizing the system for extreme environments.
DOI Requested