| The human face is a medium for conveying emotions.In daily communication,one can roughly guess what a person is thinking based on their expressions.With the development of technology and the emergence of face-related algorithms,it has become possible to capture face information through ordinary cameras,followed by the emergence of various applications.Virtual image is one of them,and it is widely used in new social media.Brilliant in the middle.However,making an avatar requires a lot of cost.It is difficult for ordinary users to own and drive their own avatars.The purpose of this article is to provide ordinary users with a low-cost way to generate avatars and can Perform style selection and secondary customization of facial features for avatars.In terms of driving,only ordinary cameras can be used to control avatars,allowing ordinary users to obtain their own avatars and use them to socialize at a low cost.There are three main technologies for automatic generation of avatars:the first is the method of using 3D scanning equipment,which requires expensive equipment support and is not suitable for ordinary users;the second is based on model library matching,based on facial features Select the appropriate model from the model library.This method has a high degree of homogeneity and small differences in roles,and cannot fit the user’s facial features well;the third method is to generate a 3D model from face pictures through deep learning.Although this method can generate an avatar that is very similar to the user,it cannot perform secondary customization and cannot choose a style.Aiming at the above problems,this paper implements a method to fit human faces according to the existing basic model.Firstly,according to the user’s face shape and the selected style,an appropriate basic model is selected,then the basic model is deformed and fitted according to the face information,and finally different rendering styles are realized by using a specific rendering method.The final test results show that the system implemented in this paper can generate a stylized virtual image similar to the user’s real face image,and has strong customization in facial features and stylized performance,and can use The ordinary camera controls the avatar,which completely realizes the whole process of stylized avatar generation and control. |
