In this short paper, we present a new approach to computer-assisted facial caricaturisation. Instead of the explicit specification of displacements, users can assign ‘virtual’ physical and material properties to the key features in an intuitive manner. The system decribed in this paper relies on only the input of facial images to be caricaturised. We examine and compare three graphical methods, namely the empirical linear model, force-motion modeland pseudo stress-strain model , for modeling and rendering exaggerated facial transformations. The use of physcal models is not intended for creating physically accurate photorealism. Our objective is instead to provide users with intuitive metaphors for specifying a desired caricaturisation,by associating virtual displacements with common physical phenomena.



   We use the multi-resolution image warping algorithm by [LCSW95 ] because it is a point-based warping algorithm, and offers a good balance between qual-ity of warping and computational costs.