KANSEI facial caricaturing based on the eye-camera interface from gallery

Facial caricaturing is a representation process of human visual impression onto a paper or other media. We have proposed a typical KANSEI machine vision system to generate the facial caricature so far This system is called PICASSO. Facial caricaturing process should be investigated and modeled from multiple viewpoints of three relations among the model, the caricaturist and the gallery. Furthermore, some kinds of interactive mechanism should be installed in order especially to implement the relation between PICASSO and gallery. In this paper, we propose a dynamic caricaturing system by using KANSEI visual information acquired from a camera. This system follows a viewer's gaze, extracts measures of the fixation and saccade, and so determines which parts of a face should be most emphasised in a good caricature. Therefore, this system could make the caricature more suitable by utilizing the viewer's own KANSEI visual information.