Huajun She1, Bian Li1, Robert Lenkinski1, and Elena Vinogradov1
This work investigates accelerating CEST imaging. The original blind compressive sensing method assumes that a few functions are enough to
represent the dynamic behavior, and the coefficient matrix should be sparse. In CEST imaging, z-spectrum shows group sparsity in the same compartment. So not only the coefficients
matrix is sparse but also the transformation of the coefficients matrix is
sparse, such as total variation and wavelets. The proposed
method addresses this prior information and further improves the original BCS method, demonstrating a better estimation of the CEST effect at high
reduction factors for both Cartesian and radial sampling patterns.