Meeting Banner
Abstract #3802

Blind Compressive Sensing Dynamic MRI with Sparse Dictionaries

Sajan Goud Lingala1, Mathews Jacob1

1The University of Iowa, Iowa city, IA, United States

We propose an sparse blind compressive algorithm to learn dictionary atoms that are constrained to be sparse for accelerated dynamic MRI reconstruction. The sparsity promoting norm on the dictionary atoms penalizes the learning of noisy basis functions. We demonstrate through examples on free breathing cardiac data, that the proposed scheme results in superior image quality in comparison to the conventional blind CS scheme and methods with fixed dictionaries.

Keywords

accelerations acquisition address alias applications arrows artifacts assume assumed atoms attenuating basis better blind blur blurring breathing capture cardiac challenge city coefficients coil coils combination compressive compromises compromising considered consistency constraining constraint constraints contains content continuation contrast convergence convex coverage dictionaries dictionary dynamic effectively error evaluate exhibit fidelity fixed flash frame frames framework free fully functions good gradient hence impose improved increment infinity inter introduced involves iterate jointly largely learn learned learning like linear local loss maintained make mask matrix minimization minimize model motion myocardial nature noise noisy observe oscillations patterns perfusion posed problem produced product profile proposed quadratic radial rank rays read reconstruction reconstructions recovery reduced reduces resolution respectively retrospective review risk robust rows sampling saturation scheme schemes sensing sensitivity series several shrinkage simple slices solution solve solved sparse sparsity spatial specifically speckle starting steps strategy subset suffered superior temporal tends term terms third trajectory uniform update utilizes variable yellow