Design And Implementation Of Sight Tracking Algorithm Based On Image Processing Technology

Eye tracking technique is the research subject which is combined with image processing and computer vision technology. The eye tracking is to use a camera and computer to realize the objective of eye watch on real time identification and tracking. With the rapid development of computer technology, Using eye tracking technology related products have been widely applied in the lives of people with disabilities, product advertising market influence analysis, visual entertainment games and so on. The main purpose of the thesis:designing a set of eye tracking the human-computer interaction system in which user does not need to match any auxiliary equipment. Besides, realizing the real-time tracking and dynamic display of the human eye fixation point on screen. Specific work of the thesis:Using the opencv computer vision library in Visual Studio 2010 development environment, build software platform of the system. Using USB camera and one of general configuration of the PC build the system hardware platform.Algorithm design is divided into face recognition, eye recognition, dynamic tracking eye pupil coordinate, The human eye canthus outside coordinates detection, building the line of sight estimation model, MFC man-machine interface design and other major parts. Firstly, based on Harr feature’s classifying algorithm detect the dynamic human face and eye, and extract the interesting part of the eye area. Then real-time tracking the pupil center coordinates by target tracking based on template matching method in the eye region. And using the improved SUSAN corner point detection algorithm locates the eyes of the outside corner of the eye coordinates, which combined with the structure characteristics of human eyes to build the line of sight estimation model, characteristic parameters with the computer screen to establish human eyes watch area mapping relationship. Finally, design system of interactive interface MFC realize the dynamic display of the human eye gaze area.Experiments show that, the thesis design the eye tracking system under the condition of natural light, to be able to more accurately estimate the line of sight watching on the screen of the area, and has low hardware requirements, users do not need to supplement the auxiliary equipment and the advantages of simple operation.