Intelligent reconstruction method of MR images

The magnetic resonance image is essential data to be processed in the neuro-imaging technique. Since the MR signal is obtained by oscillating magnetic fields in a special RF region, the MRI extraction procedure relies on complicated digital signal processing. The reconstruction of a 3D volume brain image from raw data is usually achieved by the trial-error phase angle correction method. In this paper, a straightforward heuristic method is developed to find correct phase angles, which can be used for extracting spatial information from the frequency encoded data. This intelligent method has been implemented into an automatic procedure in the reconstruction stage. It is found that the proposed method can save the neuro-imaging analysts a significant amount of time in the total analyzing procedure. Furthermore, the noise from a ghost image is also significantly reduced in the proposed method.