Contexte et objectif

The Rayleigh distribution has been shown to be insufficient for the modeling and segmentation of the echographic regions corresponding to the myocardium. The question raised is thus: are there other, better suited models, making it possible to envisage the segmentation of a larger class of echographic images? In addition, the evaluation of segmentation algorithms, motion estimation and the registration of images in thoracic and cardiovascular imaging is a difficult problem because of the lack of reference. Digital models of realistic geometries and behaviour (breathing thorax, beating heart, fluids flowing through vascular structures) can be considered as a reference. Their development would make it possible to simultaneously acquire knowledge that could then be used in the “Spatiotemporal image analysis” research axis. Also, is it possible to evaluate in advance the impact of a radiation therapy session onto the pathological tissues? Numerical simulation is one of the very promising tracks to this aim.

Methodological perspectives