Based On Video Images Of The 3d Face Model And Identity Authentication

Human face simulation on computer has always been an activity problem in computer graphics. In recent years, many results have been achieved in 3D face construction with particular face information. However, getting so much information is expensive, and it only can be used in some particular simulation. Some other methods like using one front 2D face image or one front 2D face image and one frank 2D face image have limited the angle of the face image, it can't widely using in the realism life. So in this article we will discuss the construction and recognition of the 3D human face by picking-up two face photos from one video.First, this article introduces the application, history and status of 3D face model. And we define face feature points based on the anthropometry and the anatomy.Second, we pick-up two human face photos from one video by the Nagasaka's method, compute the degrees rotation of these two photos, and solve 3D coordinate of human face feature points by the knowledge of computer graphics and geometry. And we construct the human face model using the method of constructing Candide-3 model. For this spared mesh, we use a modification of Loop subdivision to hold the original shape of the model. The face model is more realistic after texture mapping.Finally, we define a formula of error and compare 3D coordinate of human face feature points between computed and some front face photos in order to the recognition of the 3D human. And we get a good result in our experiments.