2023/03/27 by Kee, Youngwook, Kang, MungSoo, Jeong, Seongho +1
#FOS: Electrical engineering #Image and Video Processing (eess.IV) #electronic engineering #information engineering
paper · doi:10.48550/arxiv.2303.15144
We propose a novel respiratory motion-resolved MR image reconstruction method that jointly treats multi-echo k-space raw data. Continuously acquired non-Cartesian multi-echo/multi-coil k-space data with free breathing are sorted/binned into the motion states from end-expiratory to end-inspiratory phases based on a respiratory motion signal. Temporal total variation applied to the motion state dimension of each echo is then coupled in the ℓ2 sense for joint reconstruction of the multiple echoes. Reconstructed source images of the proposed method are compared with conventional echo-by-echo motion-resolved reconstruction, and R2* of the proposed and echo-by-echo methods are compared with respect to a clinical reference. We demonstrate that inconsistency between echoes is successfully suppressed in the proposed joint reconstruction method, producing high-quality source images and R2* measurements compared to clinical reference.