Artificial Intelligence Based Methods in Medicine
by Jean-Paul Haton
The RFIA/SYCO research group of Centre de Recherche en Informatique de Nancy and INRIA-Lorraine has been involved for the past ten years in several research projects related to the application of artificial intelligence techniques in medicine. These techniques have proven to be useful for helping physicians in the three main areas: interpretation of signals and diagnosis, decision-making and planning, and man-machine interaction.
The methods involved in the various projects are related to knowledge-based reasoning, connectionist neural networks and stochastic models. Most projects are carried out in close cooperation with different departments of the University Hospital of Nancy (CHRU-Nancy).
Present and completed projects deal with the following problems:
- signal interpretation in cardiology (ECG) and in the monitoring of new born babies
- surveillance of and assistance to old persons at home, based on the inter-pretation of different physiological signals (in the framework of the Diatelic project involving several French laboratories),
- planning of radiotherapy treatments of cavum cancers, based on the expertise of expert doctors and radiologists, and using a multi-agent architecture (the CAVCAV expert system, developed with the Nancy Cancer Center: CAV)
- computer-assisted anaesthesia based on the real time interpretation of several parameters, especially EEG (with the department of Anesthésie-Réanimation of the CHRU, Prof. Meistelmann)
- experiments on IRM images for the development of biologically grounded neural models (with the department of Neuroradiologie of the CHRU, Prof. Picard)
- access to image databases for diagnosis of kidney diseases, based on classifi-cation and case-based reasoning (with Dr Casanova, Marseille and Dr Barrat, Nancy)
- data mining applied to knowledge discovery in the database of children cancers (with Prof. Sommelet),
- analysis and self-control of voice for deaf adults or people with pathological voices (the SIRENE II system, developed with Thomson and ANPEDA, the national association of non-hearing people). This research is now being extended to German and Spanish in the framework of the ISAEUS project of the European Union's Telematics/TIDE program.
The field of medicine poses fundamental questions and constitutes an important application domain for Artifiical Intelli-gence methods. Our group will maintain its effort in this area in close interaction with the CHRU of Nancy, especially in the framework of the European Health Pole, recently created in Nancy. For more information on the group: http://www.loria.fr/CRIN/equipe/rfia
Jean-Paul Haton, Frédéric Alexandre, François Charpillet, Marie-Christine Haton, or Amedeo Napoli - CRIN and INRIA-Lorraine
Tel: +33 3 83 59 20 50