Abstract
Compressed sensing magnetic resonance imaging (CS-MRI) is a hot topic in the field of medical signal processing. However, it suffers from low-quality reconstruction and long computation time. In this preliminary research, we took a study of applying exponential wavelet as a sparse representation to the conventional fast iterative shrinkage/threshold algorithm (FISTA) for the reconstruction of CSMRI scans. The proposed method was termed exponential wavelet iterative shrinkage/threshold algorithm (EWISTA). Simulation results demonstrated EWISTA was superior to existing algorithms w.r.t. reconstruction quality and computation time.