Mingdong Fan1, Brendan Eck2, Nicole Seiberlich3, Michael Martens1, and Robert Brown1
1Physics, Case Western Reserve University, Cleveland, OH, United States, 2Cardiovascular and Metabolic Sciences, Cleveland Clinic, Cleveland, OH, United States, 3Radiology, University of Michigan, Ann Arbor, Ann Arbor, MI, United States
1Physics, Case Western Reserve University, Cleveland, OH, United States, 2Cardiovascular and Metabolic Sciences, Cleveland Clinic, Cleveland, OH, United States, 3Radiology, University of Michigan, Ann Arbor, Ann Arbor, MI, United States
In this study, we propose an
end-to-end deep learning based reconstruction model for MRF that aims to address the
issue of the spatial aliasing artifacts and provide accurate reconstruction
with ultra-short MRF signals.
Figure 1. Structure schematic for the deep Learning based MRF reconstruction model. It is a modified U-Net structure, composed of 19 layers and nearly half a million parameters. The model inputs the entire spatial-temporal MRF signal and directly outputs a tissue parameter map.
Figure 3. Comparison of tissue parameter maps between dictionary based reconstruction, deep learning model prediction and ground truth. The residue map is defined as the absolute difference between a reconstruction map and the corresponding ground truth map.
