MRI-PET data fusion using soft computing for the automatic assessment of myocardial viability

A data fusion system for the automatic assessment of Left Ventricular (LV) myocardium viability is presented. The system fuses LV contractile function parameters extracted from tagged Magnetic Resonance Images (MRI) and glucose metabolism rate imaged by a Positron Emission Tomography (PET) scanner. A registration package, which includes 2D and 3D visualization, has been designed. The fusion system is based on Soft Computing techniques. It is a hierarchical modular network consisting of four Adaptive Inference Systems and is able to learn and adapt itself as well as integrates expert knowledge. The particular advantage of the authors' system is that the MRI and PET complement each other lending to an accurate assessment of the myocardial viability. The presented system is a valuable tool for clinical and research applications.