Intégration de connaissances et modélisation en imagerie médicale

Image processing has considerably evolved since the first works, forty years ago. The currently observable tendency is that the methods are more and more mathematically-based with an objective of generalization of the concepts for high-level image interpretation of images. Nevertheless, these methods, despite their generic formulation, are always specialized on image modality and research context in order to obtain better and more reliable results. This specialization is carried out by integrating knowledge. In medical imaging, the studied objects are anatomical and functional structures associated with a corpus of medical knowledge which can be advantageously exploited in the image processing techniques.