The four-year, ANR-funded research project LOGIMA aims at extracting high level information from images or video sequences, when the detection and recognition of structures can benefit from prior structural knowledge (such as spatial interactions). The core of our proposal relies on

1. modeling image interpretation as an abduction reasoning process and considering dynamic information in updating and belief revision processes,
2. making mathematical morphology as a central tool for merging different formalisms that were previously disconnected, thanks to its strong algebraic foundations,
3. taking into account the uncertainty and bipolarity of spatial information,
4. exploiting the potential of graph based representations and grammars to model and

to solve the computational problem of structural scene recognition in images.

The project consortium contains researchers from the three French partners Ecole Centrale Paris, LTCI/Télécom-ParisTech and Université Paris SUd 11 as well as a new partnership with TU Dresden.