Image-Driven Automatic 3D Human Face Modeling and Editing Algorithm

Image-driven 3D face modeling is a hot topic in computer graphics.This paper proposes an image driven automatic 3D human face modeling and editing algorithm,which is based on a 3D morphable face model.The 3D morphable face model is a blended shape model built through statistical learning from a 3D face database,through which a face can be represented by a low-dimensional parameter vector.Our method extracts the image features by the algorithms of face detection,face alignment and edge extraction,and then fits the 3D morphable face model with the image to reconstruct the corresponding 3D face model.Furthermore,the 3D face model can be edited by adjusting the parameter vector.The algorithm is verified by inputting 5 face pictures and reconstructing 3D face models,and then comparing the results between the real face model and the reconstructed model.The comparison results show that this method can generate realistic 3D human face and edit the face model with various styles.