Face Modeling And Motion Simulation Based On Non-rigid Deformation

With the gradual maturity and rapid development of virtual reality technology,it involves virtual anchors,virtual shopping,virtual training,3D games,and 3D movies,which have broaden application fields and potential value.And 3D virtual face modeling and simulation of facial action expressions have become more and more enthusiastic for researchers.Many advanced technologies and methods have also appeared one after another,but usually require complex and expensive equipment and complex algorithms.In this paper,after in-depth study of key technologies such as face modeling,image processing technology,facial expression simulation,etc.,a non-rigid face reconstruction based on a single frontal image is realized,and real-time simulation can be performed according to the face movement in the video stream.The system is easy to operate,has high operating efficiency and good stability,and meets the realistic needs in human-computer interaction.The main work of this article includes:1.Research and implement 3D face reconstruction based on inverse distance weighted interpolation algorithm.First,personalized 3D face reconstruction was achieved by detecting feature points on a single frontal face image and an inverse distance weighting algorithm.Second,according to the unique topological relationship between the hole features in the face model and the model,this paper uses the floyd algorithm to solve the shortest path based on the model surface.The shortest path is used as the weight to calculate the 3D face reconstruction.Finally,the texture coordinates are calculated from the face image and the three-dimensional model.In this way,a personalized 3D face model is generated.2.The expression migration of the topology-independent face model was researched and implemented;the difficulties caused by the MPEG-4 standard when setting FAP parameters for complex models were solved;and several basic expressions of the face were simulated.First,the facial expression simulation is performed by the MPEG-4 standard.Second,cut the three-dimensional face model.A proxy model is generated.Finally,a proxy model of the same topology is used as an intermediary to transfer facial expression parameters,and a topology-independent facial expression migration algorithm is implemented,and facial expression simulation is implemented on a refined personalized face model.3.Research and implement a three-dimensional face emotional interaction system.Real-time simulation of facial expressions in 3D virtual human face is realized through the tracking and recognition of human faces.First,pose recognition is performed on the face image of each frame in the video.The two-dimensional face pose is mapped into three-dimensional space by calculating the rotation and translation matrix.Second,global and local features are extracted from facial expressions in the image.The facial expression is obtained through the classification of multi-level support vector machines.And it is displayed in real time in the 3D model.All above improve the experience of human-computer interaction.