REAL-TIME DYNAMIC DROWSINESS DETECTION USING CONVOLUTIONAL NEURAL NETWORKS

Abstract: Driver fatigue and reckless driving are major contributors to road accidents, resulting in the loss of precious lives and compromising road traffic safety. Effective and precise solutions to detect driver drowsiness are crucial in preventing accidents and enhancing road safety. Numerous driver drowsiness detection systems have been developed using diverse technologies, each focused on detecting specific parameters related to the driver's tiredness.This research proposes a novel multi-level distribution model for detecting driver drowsiness, employing Convolutional Neural Networks (CNN) technology. The model utilizes a 2D Convolutional Neural Network to analyze the driver's facial patterns, capturing their behavior and emotions accurately. OpenCV is employed to build the suggested model, and the experimental results demonstrate its superior efficiency in recognizing the driver's emotions and level of tiredness compared to existing technologies.  

Keywords: ReLu, Voila Jones Algorithm, Support Vector Machine, Convolution Neural Network, Haar Cascade, OpenCV, Keras, TensorFlow