NPAR by Example: Line Drawing Facial Animation from Photographs

This paper presents a new method for generating non-photorealistic (NPR) facial expression animations. The method consists of two main processing parts. First, a particular NPR style portrait is created from a frontal facial photograph. Facial expressive details, like: expressive wrinkles, fine, creases, even freckles, can be included in the synthesized portrait, which makes it more expressive. Second, with two different facial expression portraits, a metamorphing algorithm using distance transform is utilized to produce the 2D animation. Our morphing process is fully automated, no control point or any manual work is required. Finally, a continuous facial expression animation is produced. The resulted animation is useful in a number of potential applications, ranging from entertainment and education to teleconferencing and psychology research