Inverse diffusion method for eliminating Gibbs annular artifact of magnetic resonance image

Abstract

The invention discloses an inverse dispersion method for eliminating a Gibbs annular artifact of a magnetic resonance image. The method comprises the following steps of: reading an MR artifact image acquired through magnetic resonance imaging; calculating the gradient of each pixel from the image data and inversely diffusing the grey values of all pixels along the gradient direction; forming limiting conditions of grey value diffusion so as to make the whole diffusion process capable of adaptively adjusting the diffusion speed and to guarantee the stability of the whole diffusion process; obtaining the directional diffuseness of a diffusion equation which progresses the grey value; and updating the grey values of the pixels to acquire the corrected MR image. A high-quality image can be acquired by performing data correction on the artifact generated by cutting of the k space data.

Description

technical field

[0001] The invention relates to a post-processing method of MR images. In particular, it relates to an inverse diffusion method for eliminating Gibbs ring artifacts in magnetic resonance images. Background technique

[0002] Magnetic resonance imaging plays an increasingly important role in clinical diagnosis because of its non-radiation, multi-parameter imaging, rich tissue contrast characteristics, and sensitivity to early lesions. However, it takes a long time to scan. In clinical applications, The data sampling time is often shortened by reducing the number of phase encodings under the premise of maintaining an appropriate signal-to-noise ratio, so some high-frequency data will be lost. Gibbs ring artifacts are introduced when Fourier reconstruction is performed with only limited frequency range k-space data. The data in the central region (low frequency region) of K space determines the contrast of the MR image, while the discarded part of the edge reg...

Images