Melissa W Haskell (Athinoula A. Martinos Center for Biomedical Imaging – MGH Harvard University, Boston) will visit our lab on 10.07.2018 and give a talk entitled “Retrospective correction of patient motion in magnetic resonance imaging using reduced modeling and convolutional neural networks” at 2 PM in seminar room 01.134.

Abstract: Patient movement during magnetic resonance imaging (MRI) exams remains a large clinical challenge, due to motion induced signal corruption that occurs during the long scan time required to perform Fourier encoding. In this talk, I will present on a new software-based method to retrospectively detect and correct for these motion artifacts. The search to find the patient motion presents a large, non-linear, and non-convex optimization, which has motivated the use of reduced modeling and convolutional neural networks to tackle this clinically relevant image reconstruction problem.