Application And Research Of Eye Tracking Technology In Visual Fatigue Detection Method

As a concept that has often been discussed in recent decades,with the development of electronic information technology,wide use of electronic equipment,the probability of visual fatigue become more frequent in modern market environment filled with various electronic screen.Lots of users continue work or play in a state of visual fatigue,not only can reduce work efficiency,physical damage,but also may cause the visual system performance degradation and more injury.Therefore,it is an important problem to use a simple and effective way to test visual fatigue.The essence of visual fatigue is the reduction of visual system working efficiency.As for this problem,Letter-Search method proposed by Der-Song Lee has excellent effect.During the process of this method,users need to search for the special word in the given text.And the working efficiency of users' visual system in different condition or environment will be determined according to the final searching accuracy.However,although in most cases this method has good ability to reflect the visual fatigue condition,users can improving the searching accuracy by increasing searching time or slowing searching speed aiming at the only feature.It may cause the reduction of result accuracy.It is necessary to add other related visual features for further perfect of this method.Based on the comprehension above,this paper makes a series of improvements to Letter-Search method by combining eye tracking technology.The main contributions of this paper are as follows:(1)We propose a visual fatigue detection process based on Letter-Search method combining with eye tracking feature.In this process,we add a reading comprehension task to supply eye tracking data.As for visual fatigue condition,the accuracy of special words judgement,area of interest,pupil diameter and reading rate are the main feature to judge.Compared with Letter-Search method,the proposed approach combines with eye tracking data to analyse visual fatigue,and avoids the boundedness of original method.The form of detection process is friendly to users,and reduce the cost of detection and impact to users' visual system effectively.It also provides many experimental ideas and reference for the current viusal researching work.(2)Aiming at the vision fatigue detection process based on eye movement data,a prototype system of visual fatigue detection was designed.This system combines with eye tracking technology and provide users with a complete visual fatigue detection process.In the process of user's task execution,the system automatically captures eye movements,such as pupil data,interest domain distribution,blink frequency and so on,and takes it as the basis of visual fatigue judgment.BP neural network is used to classify experiment data and helps conclude whether the user is in visual fatigue condition or not.The BP neural network algorithm is provided by the open source code of Weka.According to the results of visual fatigue detection system,the accuracy of visual fatigue judgment reached 95.2%,which confirmed the feasibility and validity.Its performance is superior to the method based on reading efficiency defined by Di,and its accuracy rate is better than the traditional method of PERCLOS visual fatigue detection for distinguishing visual fatigue and normal state.