Expressive rendering allows users and artists to use computer graphics tools to produce their work. When dealing with animation, a wide variety of style cannot by applied using static frames techniques because of temporal coherence. Temporal coherence driven by the fact that human perception is disturb by abrupt and localized variation in an animation. So we want to control the amount and the speed of these variation of the rendering to convey efficiently users willing. This is the main subject I have work on during my PhD.
We look particularly around stroke based rendering and painterly rendering. Here each individual stroke is more or less percieved, but the image is considered to be build with a large amount of individual stroke element (contrary to continuous tone style like usually watercolor).

Digital halftoning is a process consiting of reducing the input high dynamic range to the output system range (for example a printer with two output values) in a manner that conserve the perceptual quality of the input data. I have done, during a Explora doc internship at the univseristy of Montréal, research work on ordered dispersed dither halftoning. We propose new method based on tiling of the image plane with a large (more than 1000) d ither matrix. These matrix are optimized to produce blue noise halftoning for all input gray levels.

During my 2nd year (DEA) master degree, I've work under the supervision of Joëlle Thollot and F. X. Sillion. We tried to identify rules to select features lines on a mesh. We compute a set of caracterstic line to depict object's shape (contours, rigde/valley and suggestive contours) and allow the user to modify the set. Then we interpret this set of lines and makes some statistic on the selection rate by a set of users. It was part of the ARC Miro in coloboration with Iparla team of the Labri laboratory. related publications

During my 1st year (maîtrise) master degree, I've worked under the supervision of Eric Galin on natural scene modelisation and simulation. We focus on interactive mushroom modelisation and propose a interactive tool to quickly models a mushroom family. The user provides some basic shape and information and he has some controler to introduce variation in the produced shapes. In the whole utility he can provide importance maps to place mushroom on a ground.