Abstract:
Affordable Magnetic Resonance
Magnetic resonance imaging (MRI) is well known as a flexible and powerful non-invasive diagnostic imaging technique available to clinical medicine. Higher static magnetic fields have been steadily pursued, at the expense of $1 million per Tesla. The high cost leads to very limited access. In contrast, magnetic resonance relaxometry studies the time domain signal, and can be performed in low and inhomogeneous magnetic field, providing an opportunity for low cost portable MR. A unilateral handheld scanner is extremely cost effective and could be widely used for point-of-care diagnosis on near surface organs. Challenges of such system include low signal-to-noise ratio and complicated signal modelling. I will present our recent work employing artificial neural networks in magnetic resonance relaxometry.