Unifying Compressed-Sensing Reconstruction Framework for Multidimensional MRI: Combining Novel Dictionary Models with Frame-Based Sparsity and Flexible Undersampling Schemes

Keywords

acquisition addition advanced allowing angular applications applied approaches around atom atoms best biomedical brain capturing cardiac chosen city clinical coefficients coherence combination combining complex component compressed comprises computing concise constant constraint constraints consuming contributions crossing curves define denotes desirable dictionaries dictionary diffusion dimension domain dynamic either element employ employed enable enforce enforcing example explicitly exploited exploiting extension faster fits fitting fixed flexible formulating frame frames framework free fully future gives gradient heart highly identity ignore improved incorporates intensities invariant investigating involves just kinds lake learning mathematical model modeling models motivating multidimensional multiplexed neural noise noisy novel operator optimization patch pattern patterns perfusion pixels positive principles problem propose proposed rank reconstructed reconstructing reconstruction regard regularity related representing resolution respectively rotated rotation salt sampled schemes scientific sensing similarly simulated space sparse sparsity spatial strategies strongly structure subsequent symposium synthesis temporal tight tiny tracts transform typically unified unifying validation variance variety wavelet wavelets zooming