Cognitive Ability Assessment And Eye Movement Biometric Recognition Based On Visual Searching Task

Cognitive ability refers to people's comprehensive ability to extract,process and store information.The cognitive ability tests,which are used to assess the different information processing ability of different person,have played an important role in modern education,career planning and the selection of special talents.Due to the late start of modern education in China,the scientific cognitive tests which marked by the western IQ tests were not developed in old China.At present,most of the IQ tests used in our country are based on the existing western IQ tests.With the progress of modern science and technology,we have put more and more attention on the brain research.And the cognitive ability tests based on the modern intelligent devices are also gradually developed,which will lead to a more comprehensive evaluation method of cognitive ability.The evaluation results will include not only the final results,but also the relevant parameters of the information processing.With the development of brain science,we will know more about the nature of cognitive ability,and the evaluation method will become more scientific and precise.In this paper,we conducted a study on people's information search capability evaluation based on the newly designed digital searching task.We not only studied the behavioral characteristic parameters of the subjects in the digital search task,but also studied his cognitive processing strategies through eye movement parameters.The'following conclusions are obtained through this research.Firstly,the results showed that the behavioral indicators of the subjects were positively distributed,which indicated thatthe digital searching task could be used for the evaluation of the cognitive abilities of different subjects.Secondly,the subjects' reaction time is inversely proportional to the saccadic distance,fixation time,the number of fixation points.And the saccadic distance is a reflection of the subjects' search pattern,the fixation time is a reflection of the information processing speed.This result indicates that the subjects' reaction time of this task correctly represents their visual searching cognitive ability.Thirdly,the maximum reaction time,which is a tradeoff value of subjects' reaction time and accuracy,has a bigger negative correlation with saccadic distance,indicating that it is a better behavior index of the subjects' searching strategy.Finally,we found that the answer type of the digital searching task were correlated with the correlation coefficient of the subjects'reaction time and eye movement characters.Therefore,in order to achieve a more equitable test,the different digital searching questions needs to be equalized.Compared with traditional human certification methods,biometric is harder to be duplicated,falsified or lost.The demand of the identity authentication is everywhere especially in this digital world.And with the development of science and the evolution of biometric identification technology,the biometric identification technology will be used more widely in the future.Biometric identification is mainly divided into two kinds,one is the recognition method based on physiological characteristics and the other is the recognition method based on behavioral characteristics.The first recognition methods have been well developed in the past,and some of them have already been applied in our daily lives,such as fingerprint identification,facial recognition,iris recognition,etc.Although physical features are harder to forge and replicate than traditional identification methods,they can still be duplicated and fabricated with the support of high-tech equipment.Because physiological characteristics are very stable and can't be changed by ourselves,so it will be very dangerous and troublesome if a person's physiological characteristics are stolen.Unlike physical characteristics,it is impossible to replicate an identical behavior pattern under existing technical conditions,even if you get a person's exact behavioral data.In addition,the biometric methods based on behavior characteristics will be the only choice in some scenarios with specific limitations.Therefore,the identification method based on behavioral characteristics has been paid more attention in nowadays.As the network has played an important role in our lives,many of the actions we interact with it have been used in the study of biometric identification,such as keystroke dynamics,screen touch behaviors,etc.In addition to hand movements,eye movements are also an important way for us to interact with the online world.The international research of eye movement biometric was started about ten years ago.Comparatively,there are few studies on the eye movement biometric in our country.We put forward a new kind of visual tasks,and a new kind of eye movement feature extraction method based on the problems in the research field of eye movement biometrics,such as the recognition accuracy rates,the template aging effect and the eye tracker's precision.We also calculated the results of three existing eye movement biometric methods based on our experimental data to get a comprehensive evaluation of the influence of our visual task and feature extraction method on the eye movement biometric.The results showed that the identification rates of these three biometric methods had been improved based on the datasets of the new visual task,indicating that the visual task can induce more complex eye movement signals which contains more personal identity information which will be conducive to eye movement biometric identification.In addition,through the results comparison of these four different eye movement biometric method,we found that the new eye movement biometric method had a relatively higher identification rates,and are less influenced by template aging effect and eye tracker's precision,indicating that it is a better eye movement biometric method.Finally,we did a research on the fusion methods of different eye movement characteristics,and found that it is more likely to get a better result from the fusion methods of same kinds of eye movement characteristics.However,if the two eye movement biometric methods both have a relatively higher identifications,a better result can also be obtained from the fusion method even if they are different kinds of eye movement characteristics.