A Novel Approach for Semi-automatic Construction of 3d Facial Model from Single Orthogonal Front Image of a Human Face

摘要

3D human face models have been widely used in applications such as facial animation, facialexpression recognition, face recognition, etc. Most of the techniques for constructing 3D facial model from faceimage use two orthogonal views of human face. This paper describes an easy method to generate 3D facialmodel from a single orthogonal front view of a face. The user has to manually detect the position of both eyesby marking the pupil in the single front image of the given face. Then using the distance between the two eyesand the specified ratio of the generic facial geometry, face boundary and other facial features are detected. Thenthe facial features and the generic 3D face model are used to generate the 3D facial model.