INTELLIGENT DRIVER DROWSINESS DETECTION USING GAOPTIMIZED CNN ARCHITECTURE

Abstract

The implementation of real-time detection technology is essential, as the danger of driving while fatigued is a significant global concern that results in a significant number of accidents. This article delineates a more advanced model for the detection of fatigued drivers by utilizing a genetic algorithm (GA) to improve a convolutional neural network (CNN). In order to enhance accuracy and reduce false alarms, the GA modifies critical CNN parameters, including the learning rate, filter size, and number of layers. The system is capable of accurately identifying when a motorist is fatigued by observing their facial expressions and tracing their eye and head movements. Our method is more precise and responsive than traditional CNN models, as evidenced by experiments. The landscape of transportation safety could be completely transformed by this CNN model modified for GA, which could reduce the occurrence of accidents and save lives. This is a result of its reliability and scalability.