Eye Accessing Cues Based On Pupil And Eye Corner Location

The theory of eye accessing cues suggests that when people think about the problem,the eye position is related to the type of thinking.In this paper,we design a set of eye clues interpretation model system,through computer image processing technology to speculate on the type of thinking.This system is mainly divided into three parts:image acquisition part,eye feature extraction part and eyeball clues interpretation model.In this paper,the head mounted image acquisition device based on infrared camera,the image can be collected to ensure that the pupil characteristics are significant,and the image is stable,easy to extract the eye feature information.In this paper,we introduce the eye feature extraction algorithm and the eye cue interpretation model.Eye feature extraction contains two parts:accurate positioning of the pupil and accurate positioning of the eye corner.In order to detect the gaze direction,accurate location of the pupil position is the basis.Therefore,this paper proposes the way of pupil localization which combines the gradient of the rough positioning based on image pyramid with ellipse fitting accurate positioning based on the least-square approach.The new algorithm for pupil localization proposed in this paper is based on the method of hierarchical detection which draws on the advantages of fast and high accuracy of the gradient detection algorithm.The range of the pupil is determined by the distance between the pupil of the eye and face,and therefore the paper proposes the eye location algorithm based on dynamic threshold and contour search.In this paper,the corner of the eye location algorithm makes use of head mounted infrared camera which can make some gray differences between the characteristics.The establishment of the interpretation model of eye cues is the basis for the correct analysis of thinking types.In this paper,through the study of the internal structure of the eye and the eye movement pattern,we can determine the model of eye clues.Then,according to the content of the theory,the eye region is divided into two parts,and then the eye cue is determined according to the vector relationship between pupil and pupil.Finally,the paper builds a eye unscrambling system,through the VS2010 development environment,combined with the Qt interface framework library functions and Opencv image processing library function,using the C++ programming language for software development.In this paper,the pupil localization experiment is used to verify the accuracy and the time of the algorithm.In addition,through the design of the system,we have carried out the experiment of eyeball clues to verify the feasibility of the software.