A Novel Fast Dynamic Cardiac Data Reconstruction Method Using Prior Knowledge and Adaptive Matching Pursuits

Muhammad Usman¹, Philip G. Batchelor¹

Recently, variants of matching pursuit methods such as Regularized Orthogonal Matching Pursuit (ROMP), Compressive Sampling Matching Pursuit (CoSAMP)and Sparsity Adaptive Matching Pursuit (SAMP) have been proposed and shown to give exact reconstruction for sufficiently sparse signals. Compared to the conventional OMP, these variants offer faster reconstructions. The SAMP has an additional advantage that it does not require any prior knowledge about the signal sparsity which is the case when recovering the practical compressible signals. Based on SAMP algorithm and the prior knowledge obtained from the sliding window reconstruction from the under-sampled data, we propose a scheme that can provide faster dynamic cardiac MR reconstructions compared to exisiting iterative schemes such as OMP.