›› Goal, motivation, history

In order to face the drawbacks of cardiac monitoring systems, two labs from IRISA and LTSI launched a common project, named Calicot, in 1999. The DREAM lab from IRISA had a strong experience in model-based diagnosis. The LTSI had investigated signal processing and cardiovascular diseases for many years. Precisely, the LTSI was at the head of the european project "Artificial Intelligence in Medecine Knowledge Based Interactive Signal monitoring System" (AIM-KISS) in 1989.

As a result Calicot has given rise to the intelligent cardiac monitoring framework IP-Calicot. This tool aims at analyzing online input signals (electrocardiograms, pressure signals) and at detecting and at characterizing cardiac arrhythmias.

The original contributions of the project are:

• to couple advanced research in signal processing and artificial intelligence in the field of diagnosis and machine learning,
• to implement an adaptive system able
  • to choose the signal processing algorithm that is the best suited to the current context,
  • to adapt online the diagnosis model to the relevant candidate pathologies.

The investigated research domains are:

• in the signal processing field:
  • the design and implementation of algorithms for detecting and classifying QRS complexes and P wave detection algorithms,
  • the conception and implementation of context estimators based on physiology, noise level, noise type...
  • the conception of strategies for selecting online the best suited signal processing algoritms according to the observed context,
• in the artificial intelligence field:
  • diagnosis methods by chronicle recognition,
  • machine learning aiming at automatic model acquisition for model-based diagnosis,
  • self-adaptive methods for piloting the system according to the context.