PRESENTATION AND OBJECTIVES :

In many engineering domains, phenomena are observed through different information supports. These supports sometimes carry complementary and independent information. In this case, taking into account simultaneously the different observations is necessary for a complete understanding of the generative process. In other cases, the different supports carry redundant information at diverse levels and studying them as a whole allows, besides a complete understanding, a greater robustness of the analysis. Extracting parameters from complementary and/or redundant information flows of heterogeneous natures can be regarded as using different modalities in the analysis process which justifies the naming of multimodal analysis or multimodal information processing.

With regard to this description, the problem the IMS group is dealing with is to extract meaningful information from these heterogeneous data and to make use of them to build systems able to reproduce (simulation), classify (detection, pattern recognition) or control (automation, robotics, human-computer interfaces) the behaviour of the systems having generated the observed information or even to enhance their quality (filtering, restoring) or to compress them. The ultimate aim of the IMS research group is to build situated systems, able to behave optimally in their environment and interact with it using their perceptions. The integration of perception (sensing), action (behaviour) and reasoning in a single process is known as the situated perception paradigm.

RESEARCH THEME :

• Statistical Signal Processing ;
  
  
• Machine Learning and Artificial Intelligence ;
  
  
• Intensive and Distributed Computing.