Font Size: a A A

Research On 3D Face Modeling And Facial Expression Synthesis Based On Single Frontal Photo

Posted on:2012-02-13Degree:MasterType:Thesis
Country:ChinaCandidate:X R XuFull Text:PDF
GTID:2218330338967548Subject:Control theory and control engineering
Abstract/Summary:PDF Full Text Request
With development of computer application technology, avatars have stepped into people's sight by many ways. The methods of realistic 3D face modeling and facial expression synthesis can be used in avatar technology to realize the personalization of avatar. Especially, if we realize the modeling and facial expression synthesis of the specific 3D face based on one picture provided by users, which can make users to have more immersion feelings. The technologies of 3D face modeling and facial expression synthesis refer to many subjects such as computer vision, computer graphic, pattern recognition, even psychology. It also can be used widely in movie effects, games, human-computer interaction systems. So it has great value in both theory and application.We can only get a single frontal face picture in many cases. So how to model realistic 3D face and imitate facial expression synthesis using a single picture is the goal of many researchers. The project studies such a system which needs to extract face feature points from the input picture firstly. We choose 34 points to extract based on Candide-3 model. After face location and image preprocessing, we extract these points by using many algorithms and characteristic of images which includes the rule of face structural distribution, local projection algorithm, and different distribution of basic colors between lip-color and skin-color. Then, we make a series modification on the vertexes of general 3D face model to generate specific 3D face model which has similar shape to input picture. After that, we use texture information from the input picture to complete texture mapping, so we can get the realistic personalized 3D face model. According to face action code system, we study facial expression synthesis of specific 3D face, and thereby users can control specific 3D face to generate corresponding facial expression by inputting predefined words. Finally, we realized above-mentioned functions by using Visual Studio 2010 and OpenGL graphics library.
Keywords/Search Tags:Single image, Feature extraction, 3D face modeling, Facial expression synthesis
PDF Full Text Request
Related items