Performance Evaluation of Optimized Convolutional Neural Network Approach in Prediction of Fatigue Driver

Abstract

The higher rate of road accidents occurred due to fatigued drivers who are the responsible for a large number of casualties. In today's environment, safe driving has gained significance. The main factors of drowsiness are inadequate rest and continuous driving on the road leads to fatigue. The fundamental motivation for designing a model that measures a driver's fatigue problem is to prevent vehicle accidents, which will benefit millions of people worldwide. The model is based on the convolutional network which predicts the fatigue level of the driver. The early prediction of drowsiness can be identified as the problem easily with the proposed model. This work finally proposes the prediction techniques by using the deep learning structure compared to the existing model. In order to predict the driver's fatigue status eye flickering, dozing, duration of eye closure and yawning frequencies without any non-intrusive approach can be predicted and classified with the help of the proposed work.