Wavelet image decomposition based variable density compressive sampling in MRI

This paper considers compressive sampling in magnetic resonance imaging. The structure of the sparsity domain is analyzed and based on the conclusions, probability profile for random sampling is proposed. This probability profile divides the k-domain into tiles with constant probability for sampling. The value of the probability for sampling k-space frequencies in every tile is found based on the content of the specific image or similar images. The performance of the proposed probability profile is compared to the performance of the distance based probability profile.