Parallel Magnetic Resonance Imaging Based On Variable Density Random Sampling Algorithm

This paper described the development process of parallel magnetic resonance imaging, the principle and the image reconstruction. Then, it introduced GRAPPA algorithm and SENSE algorithm which had been used in clinical application. At last, proposed a new variable density random sampling (VDRS) algorithm, and accomplished in the C++environment.The majority of the image energy is concentrated in the center of the low-frequency area; less energy is concentrated in the edge of the high-frequency area. The VDRS algorithm was used in this paper, which sampled the low-frequency information primary and high-frequency information rarely. Thus it can reduce the aliasing artifacts effectively. Finally, we can get high-quality images through the Fourier transform, image reconstruction and multi-coil fusion technology.The experiment result showed that, compared to GRAPPA algorithm, the VDRS algorithm scanned reconstruction image with high quality and faster speed under the same conditions. Compared to the SENSE algorithm, the VDRS algorithm reduced image aliasing artifacts greatly under the same reduction factor. The VDRS algorithm was accomplished in the C++environment, the quality of the reconstructed image was the same as in the MATLAB, but the former accelerated the speed of imaging.