Research project

Recent advances in Multimodality Imaging have shown that a PET-MRI integration is indeed feasible. This fusion capability makes it now possible to image the functional metabolism (PET) in conjunction with soft tissue contrast (MRI). However, the attenuation correction for PET images still remains a problem at large. Existing approaches to tackle this issue relies on using the CT anatomical information and registering it in a non rigid manner onto the MRI data sets in order to generate linear attenuation maps. Another approach is to assign approximate attenuation coefficients to MRI brain images using a segmentation procedure. The problem though is that MRI conventional sequences fails to image the cortical bone which in turn results in a misclassification of the bone as air, both possessing different attenuation coefficients. But Ultra-Short-Echo sequence(UTE) has shown promising results in imaging the cortical bone. Our approach relies on investigating the feasibility of deriving an MRI based attenuation correction mechanism based on UTE sequences and pattern recognition methods.