Multi-dictionary learning for magnetic resonance image reconstruction based on entropy and geometric orientation

Abstract

The invention relates to the magnetic resonance imaging technical field, and specifically relates to a classification multi-dictionary learning magnetic resonance image reconstruction method based on entropy and geometry directions; an existing DLMRI algorithm is poor in detail portions when reconstructing a nuclear magnetic resonance image; the classification multi-dictionary learning magnetic resonance image reconstruction method comprises the following steps: using a downsampling model to obtain partial K space data; building a magnetic resonance image reconstruction model for the obtained partial K space data; carrying out inverse Fourier transform for the partial K space data so as to obtain an initial image; splitting the initial image into overlapped image blocks; solving entropy of each image block, splitting the image block samples into four classes according to the entropies from small to big, further classifying the latter two image block classes according to the geometry directions, carrying out dictionary training for the image block samples, solving a sparse coefficient corresponding to the dictionary, and thus obtaining a reconstruction image matrix; carrying out Fourier transform for the reconstruction image matrix, updating the image matrix, carrying out inverse Fourier transform for the updated image matrix, thus obtaining the reconstructed magnetic resonance image.

Description

technical field [0001] The invention relates to the technical field of magnetic resonance imaging, in particular to a magnetic resonance image reconstruction method based on entropy and geometric direction for classification and multi-dictionary learning. Background technique [0002] Magnetic Resonance Imaging (MRI) technology is one of the most widely used medical imaging methods. Compressed Sensing (CS), as a new sampling theory, provides a key theoretical basis for MRI, namely On the premise that only part of the K-space data is acquired, MR images that meet the needs of clinical diagnostic quality can be obtained by optimizing the reconstruction algorithm. The premise of accurate reconstruction of MRI by CS-MRI is that the MR image can be sparse in a certain transformation domain or dictionary. Representation, the inverse problem of the image is solved with the sparse prior of the signal, so that the accurate reconstruction of the MR image can be realized by using part ...

AI Technical Summary

Problems solved by technology

[0006] The purpose of the present invention is to solve the problem of insufficient details when the existing DLMRI algorithm reconstructs nuclear magnetic resonance images, and proposes a magnetic resonance image reconstruction method based on entropy and geometric direction classification multi-dictionary learning to achieve more accurate image blocks Classification, improve the sparse ability of the dictionary, while retaining the detail components of the image to eliminate artifacts and improve the reconstruction quality of magnetic resonance images

Images