Generation of 3D facial expressions using 2D facial image

The creation and animation of 3D facial models mostly depends on animators' manual work by the frame. Thus, it needs a lot of effort and time as well as various hardware and software. This paper is aimed at introducing a facial expression modeling and caricature system which is based on muscle vector and produces diverse and realistic facial expressions. Based on the eight standard facial models for Korean faces, we suggest a way to model 3D human faces easily and quickly with the front and profile images. In addition, an elegant adjustment method which produces more precise 3D facial models using the linear spline is proposed. The proposed method can be easily applied to producing caricatures by adjusting the control points of the spline and vectors of the facial muscles.