Representation and acquisition models for expressive rendering

This thesis is structured along three research axes: First, in I investigateg the representation issues raised by the production of expressive images and animations, (that is to say, introducing the artist/user's expressivity). It led me to propose solutions to practical problems in this domain. Second, I studied the acquisition of low-level notion of styles from user-supplied examples. In particular, I produced aquisition systems working on vector-based as well as bitmap color images. Finally, I have been personally motivated by the fascinating field of human vision and its many links with expressive rendering. Thus, throughout this thesis, I tried to build as many connections as possible between the two fields, hence building a framework where various expressive rendering approaches can be studied in a principled way.