Meeting Banner
Abstract #2648

Dynamic Parallel-Imaging Reconstruction with Image-Block Dictionaries

Eric Y. Pierre1, Nicole Seiberlich1, Vidya Nadig2, Mark A. Griswold3

1Biomedical Engineering, Case Western Reserve University, Cleveland, OH, United States; 2Heart and Vascular Center, MetroHealth Campus of Case Western Reserve University, Cleveland, OH, United States; 3Radiology, Case Western Reserve University, Cleveland, OH, United States

Dictionaries with spatial or temporal information shared with a target image are often used as a sparsifying transform to improve Compressed Sensing reconstruction results. However in accordance to Parallel Imaging principles, the coil sensitivity information within these dictionaries could also be directly exploited to reconstruct undersampled images with a light computational load. A new reconstruction scheme which exploits image-block dictionaries extracted from a calibration data is proposed. In vivo reconstruction results are shown for a dynamic cardiac experiment with radial acquisition accelerated by a factor 8.

Keywords

abdominal accelerated acceleration acceptable accurate achieve acquisition adaptive addition aims aliased aliasing assembled assumption atoms audience bias biomedical block blocks bottom breath calibration campus cardiac channel clear close coil combination combined comes compensated compensation complex compressed computation computational computing conditions consecutive constraints correlated density diastolic dictionaries dictionary dimension dynamic either ejection entry evaluate even experiment exploit exploited extracted feature fidelity final fraction frame frames functions gating heart highly identical implicitly introduced involved iterative latter learning light load local location long maps mark matrix medical middle motion must needed noise novel object operations operator outside parallel partially pixels posed powerful processing produce projections proposed quality radial radiology rather real receiver reconstructed reconstruction reconstructions reduction regularization relying require required reserve resolution sample scanned scheme sense sensing sensitivity separable simple since slice solutions sparsity specifically spine step subject supported systolic target though training trajectory transactions transforms type unlike validity various vascular western wise yielding