In a first part, we propose an innovative methodology for image matching in the context of reservoir simulation. In order to build a model consistent with data collected on the field, we need to evaluate the error between seismic cubes obtained by simulation and seismic cubes acquired in the oil field. Using image processing tools, we develop a new formulation of the error. The application of this new formulation on synthetic reservoir cases demonstrates its efficiency. In a second part, we address the issue of designing two theoretically well-motivated registration models capable of handling large deformations since they are based on nonlinear elasticity. The shape to be matched are viewed as Ciarlet-Geymonat materials for the first model and as Saint-Venant Kirchhoff materials for the second one. We investigate the efficiency of the proposed matching model for the registration of mouse brain gene expression data to a neuroanatomical mouse atlas.
