Research And Implementation Of Real-Time Simulation Of Facial Expression Based On Video

Human face is the window of human emotion,and the human facial expression provides a wealth of visual information.Face facial expression simulation in the production of film and television special effects,animation and other entertainment,as well as customer service,auxiliary teaching,video conferencing,and many other fields have a wide range of applications.Real time simulation of facial expression based on video technology involves face recognition technology,face feature point automatic positioning technology,face pose estimation and facial expression analysis and facial expression animation.In this thesis,a video driven facial expression simulation system is designed and implemented.The system can locate the position of the human facial feature points in the video,and combined with the position information of the facial feature points to estimate the face pose and facial expression.Then,based on the facial expression data,the 3D facial expressions are fused to realize the realistic facial expression animation with the three-dimensional Morphing technology.The main work of this thesis is reflected in the following points:1.Face and eye detection.Adaboost learning algorithm is used to detect the human face and eyes in the video image,which provides the optimal positioning data of the initial model.2.Facial feature points automatic positioning.According to the feature of the given sample,the active shape model is used to generate the shape model and the gray texture model.This paper proposes an improved method of ASM feature point positioning,accurate positioning model of the early use of eye detection,image based on multi-resolution search strategy to enhance the precision of algorithm,combined with the application of algorithm in the video presents a search optimization method of ASM feature points based on video.3.Human face pose estimation.Based on the position information of face feature points,a linear regression model is used to estimate the deflection angle of the human face.4.Human facial expression analysis.According to the estimation of facial feature points and pose information,this thesis proposed an analysis method for facial expression facial action unit based on local facial motion analysis according to the position of distance between feature points,the facial feature points location information into facial expression of digital information,used to drive 3D model animation.5.3D facial animation.The 3D model is driven by the pose and expression information.A grid Morphing method is used to fuse the basic facial expression models to form a realistic facial expression animation.