Study Of Gaze Estimation Method Based On Face Normal

The eye gaze direction or the fixation point usually reflects what people are interested in. Therefore, the line of sight is a very good human computer interaction input channel. Fixation point estimation technology is to estimate the human eye gaze direction or position with the computer in the computer industry. In the recent years, as the human-computer interaction interface, the fixation point estimation technology has attracted extensive research at home and abroad. For a long time, many researchers have been trying to apply it in other research areas, such as ophthalmology, advertising, psychology and so on.For main problems of the existing gaze estimation methods which are that they restrict the head movement of users and the infrared light has damaging effects for eyes, this thesis presents a gaze estimation method based on face normal. Stereo vision is used in the method. First, to calibrate stereo cameras to determine the extrinsic and intrinsic parameters of camera; secondly, face is quickly recognized by using Adaboost, to locate iris based on integro-differential operators to determine the center position of the two pupils, and according to the different saturation, the center 2D coordinates of the noses and mouth are acquired; thirdly, the 3D coordinates of these five points were obtained by limit-restraint, stereo matching and 3D reconstruction; then, a plane fitting in the least squares is adopted to get the approximate face plane, and find out plane normal on midpoint of the two pupils. Finally, the point of gaze can be obtained by getting the cross point of normal and computer screen. The method can be a great aid to position a moving object.