Eye Tracking Technology Application In The Human-computer Interaction

With the progress of society and technology, intelligent equipment into the aspirations of people.Eye tracking technology is applied widely. Eye tracking technology research in human-computer interaction is a hot research topic in the computer vision, Interactive application can respond to the user’s gaze in some way. Interactive eye tracking technology is suitable for human-computer interaction, visual display, and the computer graphics field. Here is the line of sight of characteristic parameters as the input data, for the application. The study of the whole system including the face and eye detection, pupil tracking, the establishment of the line of sight direction and mouse control information generated.In this paper, the detection algorithm based on AdaBoost to test the characteristics of face and eye; Here by Haart characteristics to optimize the weak classifier for face detection, eye detection is through AdaBoost algorithm, The eye area filtering using integral projection method to further reduce the eye area, Application opening function of mathematical morphology for the pupil target type., This experiment in the four corners screen installed near infrared light source, therefore in the pupil area will can detect four s&p khin spot. First to test the pupil center, first to filtering, binarization pupil area. Then use the hough transform combined with least squares straight line fitting method to calculate the pupil center. To test the s&p khin spot location again, Here we use first Canny edge detection algorithm for s&p khin contour detection, Then the coordinates of two s&p khin spot.Through the pupil center and the s&p khin spot center coordinates can calculate the line of sight point coordinates. On the basis of the detected the pupil, pupil tracking on to the next frame. Adopted here is that the combination of particle filter and kalman filter algorithm for tracking. Performed by particle filter in the next frame image pupil goal orientation, Then the accurate segmentation pupil in this position. The division of the pupil parameters using kalman filter to estimate.The last line of sight direction estimation using a completely depend on the space geometry calculation method of eye placement operation by the user. Due to the corneal curvature center and the pupil center on the same optical axis, so the space of the optical axis direction is obtained by computing the corneal curvature center and the virtual pupil center. Second to determine the visual axis, through the calibration of a 3* 3 axis rotation matrix M indirectly obtained the relationship between the visual axis and the optical axis, to determine the visual axis direction; Finally, the intersection point of the screen and visual axis is obtained by calculation, the fixation point.Eventually we will get the fixation point applied to the screen, to complete the basic control of the mouse operation, such as the mouse left, right, open the document, input simple text. etc.