Coherence analysis of compressive sensing based magnetic resonance imaging reconstruction

This thesis focuses on the coherence analysis of compressive sensing-based magnetic resonance imaging reconstruction. One major problem of magnetic resonance imaging is its relatively long data acquisition time. To reduce data acquisition time, a approach is to acquire a portion of raw data. Besides, new encoding methods, such as noiselet encoding, were applied to replace the original Fourier encoding. The original contribution of this thesis is the first coherence analysis of the difference between Fourier encoding and noiselet encoding. The results provide guidance on design for 3-dimensional volume imaging using novel encoding techniques in magnetic resonance imaging.