The Augmented Reality System Research And Implementation Of Face Locating And Tracking

Human face is the most common mode in people’s daily life. It is also a very important channel of expressing people’s feeling and reflecting people’s individual characteristics. With the continuous development of augmented reality technology, facial augmented reality becomes an active topic in both computer vision and computer graphics, which has broad prospects of applications in entertainment industry, such as human-machine interaction, multimedia and video communication, ect. Particularly in recent years, Iphone and Ipad and other handheld devices, which have high-resolution display panel and GPS sensor, are become more and more popular. They make a solid foundation of commercial application of facial augmented reality. And facial augmented reality will become new round of boom without doubt.This thesis gives a creative solution to solve the problem of facial augmented reality. This approach first detect human face form the camera image by AdaBoost algorithm, then get the coordinates of key feature points by AAM algorithm, and construct three-dimensional human face matching mesh, then estimate the face pose parameters by POSIT algorithm, insert the virtual objects to the three-dimensional face mesh model finally, get the effect of augmented reality at last. The thesis also analyzes the facial augmented reality system’s main function modules. Base on the study of key algorithms and system’s structure, an augmented reality demo which can help people try different glasses was realized. The results demonstrate that the approaches and the system are feasible and practical.