One of our goals is to develop advanced visualization techniques and apply them to real-scale problems.We developed algorithms operating on the GPU for large voxel grids, vector fields and for implicit surfaces.
Voxel grids: The main application domain is oil and gas exploration. The challenge is to extract information from a "seismic cube", i.e. an "ultrasound image" of the underground. This is used for planning the path of the wells in order to maximize the chances of discovering oil. Our approach combines out-of-core data structures with advanced visualization. tools. We also develop distributed visualization algorithms for clusters. We also developed GPU-based algorithm, for visualizing iso-surfaces in tetrahedral grids [ISVC 2006].
Vector fields: We developed a formalism for manipulating vector fields defined over surfaces [ACM TOG 2008], [ACM TOG 2009] and a visualization method [IEEE VIS 2006] for vector fields with singularities of high order.
Implicit surfaces: We developed a ray-casting algorithm for algebraic implicit surfaces [ISVC 2007]. We applied this algorithm to visualize large CAD models with tubes and pipes [TMCE 2008],[ISVC 2008] and molecular surfaces [J. molecular graphics 2008]. This article was in the top 25 downloaded papers of Journal of Molecular Graphics, and Matthieu Chavent was awarded the national Ph.D prize from GMM Group (Graphics and Molecular Modeling). We also developed a ray-caster for height fields and used it to introduce geometry textures [Computer Graphics Forum 2008], a method to add fine details to a coarse geometry.