Research On High-speed Eye Tracking Based On Autostereoscopic Display

With the advancement of stereoscopic display technology,people will continue to pursue a better visual experience.The autostereoscopic display technology provides viewers with better comfort and flexibility because it does not require the assistance of equipment such as glasses.Among them,the raster-type autostereoscopic display has the fastest development speed and has been commercialized.However,this technology has a major flaw,that is,the position of the best viewer is fixed.When the viewer moves to another area,crosstalk will occur,which seriously affects the viewing experience.In order to solve this problem,a high-speed human eye tracking study based on autostereoscopic display is carried out in this paper.The high-speed human eye tracking algorithm is used to locate the viewer's eye position in real time,and the optimal viewing position is changed using image compensation.Finally,the human eye tracking software based on OpenCV is realized,and the human eye tracking autostereoscopic display is completed.The main research contents of this article are:1.In-depth study of the binocular vision model based on the principle of binocular vision,and introduced the raster autostereoscopic display technology based on the principle,focusing on the analysis of the problems of crosstalk and narrow viewing range faced by the technology,and finally proposed image-based compensation With the human eye tracking solution,that is,the human eye tracking autostereoscopic display system,this method can effectively reduce crosstalk and ensure that the viewer has a good viewing experience at all positions.2.Research various algorithms to implement a high-speed human eye tracking module.Firstly,I deeply researched the YOLO algorithm,introduced the principle of the YOLO algorithm and trained the YOLO network to realize face detection.Then,a random forest-based human eye positioning algorithm is introduced.The random forest is used to train local binary features for face alignment,and the algorithm is used for human eye positioning.Finally,the eye tracking algorithm based on correlation filtering is analyzed,and the principle and process of the correlation filtering tracking algorithm are analyzed.The above algorithm is combined with OpenCV to realize high-speed eye tracking software,which greatly improves the tracking speed on the basis of improving the detection accuracy,and meets the application requirements of the eye tracking free stereo display system.3.Introduced the structure of human eye tracking autostereoscopic display system.It explains the flow of the eye tracking module,introduces the measurement and use of the corresponding table in the index search module,and implements the eye tracking software running on the Windows system.Finally,the hardware implementation of each module is introduced,and the actual objects and the overall effect of the system are displayed.