2021/11/21 by Qing Zou, Abdul Haseeb Ahmed, Zou, Qing +9
Medicine · Physics and Astronomy · #Advanced MRI Techniques and Applications #Atomic and Subatomic Physics Research #Computer Vision and Pattern Recognition (cs.CV) #FOS: Computer and information sciences #FOS: Electrical engineering #Image and Video Processing (eess.IV) #Medical Imaging Techniques and Applications #electronic engineering #information engineering
paper · pdf · doi:10.48550/arxiv.2111.10889
openalex publication_date 2021/11/21 · openalex created_date 2022/11/01 · openalex updated_date 2026/07/28
Free-breathing cardiac MRI schemes are emerging as competitive alternatives\nto breath-held cine MRI protocols, enabling applicability to pediatric and\nother population groups that cannot hold their breath. Because the data from\nthe slices are acquired sequentially, the cardiac/respiratory motion patterns\nmay be different for each slice; current free-breathing approaches perform\nindependent recovery of each slice. In addition to not being able to exploit\nthe inter-slice redundancies, manual intervention or sophisticated\npost-processing methods are needed to align the images post-recovery for\nquantification. To overcome these challenges, we propose an unsupervised\nvariational deep manifold learning scheme for the joint alignment and\nreconstruction of multislice dynamic MRI. The proposed scheme jointly learns\nthe parameters of the deep network as well as the latent vectors for each\nslice, which capture the motion-induced dynamic variations, from the k-t space\ndata of the specific subject. The variational framework minimizes the\nnon-uniqueness in the representation, thus offering improved alignment and\nreconstructions.\n