A New Method Of MRI Image Reconstruction Based On All Phase FFT

Magnetic Resonance Imaging(MRI) is one of the most remarkable medical imaging technologies in recent years. MRI is not only non-invasive to human body, but also can generate image from almost any direction and give details on soft-tissue structures. These benefits make MRI increasingly widely used in practice.However, conventional MRI is time-consuming, and may generate artifacts in the reconstructed images. In clinical application, high time-efficiency is often obtained by reducing the scanning data, i.e. partial k-space imaging. Directly using Fourier inversion on partial k-space data can lead to Gibbs phenomenon, which will lower the image quality and cause misdiagnosis. Therefore, it is particularly important to study how to eliminate the truncation artifacts in partial k-space imaging.All-phase Fourier reconstruction is a combination of all-phase theory and Fourier transform, targeting on discontinuous signal reconstruction. It is currently used mainly in 1-D signal reconstruction. This thesis extends this method to 2-D MRI image field to eliminate the truncation artifacts caused by partial k-space imaging in MRI. When using all-phase FFT method to reconstruct the partial k-space data, firstly the original data is extended through circular shift operations, secondly the extended coefficients are multiplied by corresponding high-frequency harmonics. This method can reduce the reconstruction error due to data truncation and remove the Gibbs artifacts caused by partial k-space imaging.This thesis uses all-phase FFT method to reconstruct the partial k-space data. Compared with the traditional zero-fill reconstruction method, all-phase FFT can eliminate the artifacts effectively and preserve the high resolution of MRI images. In conclusion, all-phase FFT method is promising in eliminating the Gibbs phenomenon in image reconstruction.