DEFINING SUB-REGIONS IN LOCALLY SPARSIFIED COMPRESSIVE SENSING MRI

Fuleah A.Razzaq, Shady Mohamed, Asim Bhatti, Saeid Nahavandi

Keywords

Magnetic Resonance Imaging, Compressive Sensing,Sparse Signals, Fourier Transform, Signal-to Noise Ratio(SNR), L1 Minimization.

Abstract

Magnetic Resonance Imaging (MRI) is an important imag- ing techniques. However, it is a time-consuming process. The aim of this study is to make the imaging process ef- ficient. MR images are sparse in the sensing domain and Compressive Sensing exploits this sparsity. Locally sparsi- fied Compressed Sensing is a specialized case of CS which sub-divides the image and sparsifies each region separately; later samples are taken based on sparsity level in that re- gion. In this paper, a new structured approach is presented for defining the size and locality of sub-regions in im- age. Experiments were done on the regions defined by pro- posed framework and local sparsity constraints were used to achieve high sparsity level and to reduce the sample set. Experimental results and their comparison with global CS is presented in the paper.

Important Links:



Go Back