CN-115345953-B - DWI-EPI image Nyquist ghost correction method

CN115345953BCN 115345953 BCN115345953 BCN 115345953BCN-115345953-B

Abstract

The invention relates to a DWI-EPI image Nyquist ghost correction method which comprises the following steps of obtaining magnetic resonance original data, dividing the magnetic resonance original data, extracting reference scanning data for reference scanning correction, filling corresponding k space positions with residual data to form k space data in a (Kx, ky) domain, converting the k space data in the (Kx, ky) domain and the reference scanning data into (x, ky) mixed domain data, performing reference scanning correction on the (x, ky) mixed domain data to obtain first correction data, converting the first correction data in the (x, ky) mixed domain into (Kx, ky) domain data, performing non-reference scanning correction on the (Kx, ky) domain data to obtain second correction data, performing zero filling processing on the second correction data to obtain third correction data, and performing two-dimensional inverse Fourier transformation on the third correction data to obtain a non-artifact image. Compared with the prior art, the invention has the advantages of good correction effect, short correction time and the like.

Inventors

CAI XIN
DENG XIANMIN
NIE SHENGDONG
LI JIANQI
HOU XUEWEN
JIANG XIAOPING

Assignees

上海康达卡勒幅医疗科技有限公司

Dates

Publication Date: 20260505
Application Date: 20220817

Claims (9)

1. A method for correcting nyquist ghost of a DWI-EPI image, comprising the steps of: acquiring magnetic resonance original data, dividing the magnetic resonance original data, extracting reference scanning data for reference scanning correction, filling the residual data in corresponding k space positions, and forming k space data in a (Kx, ky) domain; Transforming both the k-space data and the reference scan data in the (Kx, ky) domain into (x, ky) hybrid domain data; Performing reference scan correction on the (x, ky) mixed domain data to obtain first correction data; Transforming the first correction data in the (x, ky) mixed domain into (Kx, ky) domain data; Performing non-reference scanning correction method correction on the (Kx, ky) domain data to obtain second correction data; Performing zero padding processing on the second correction data to obtain third correction data, and performing two-dimensional inverse Fourier transform on the third correction data to obtain an artifact-free image; the segmentation of the magnetic resonance raw data is specifically as follows: According to the sequence gradient information, the acquisition frequency and the system delay information, discarding the signal data acquired at noise points and read gradient climbing positions, only taking data points in a read gradient steady section in cyclic acquisition, and removing non-phase coding data, so that the data matrix size is consistent with the matrix size set during scanning.
2. The DWI-EPI image nyquist ghost correction method according to claim 1, wherein the reference scan data is data with ky=0 in a scan-acquired data matrix.
3. The DWI-EPI image nyquist ghost correction method according to claim 1, wherein the negative gradient data in the k-space data is inverted left-right so that the k-space data is uniform in the more directions when filling the corresponding k-space positions.
4. The DWI-EPI image nyquist ghost correction method according to claim 1, characterized in that the inverse fourier transformation of the k-space data and reference scan data in the Kx direction, respectively, transforms each differently phase encoded signal into the (x, ky) hybrid domain.
5. The DWI-EPI image nyquist ghost correction method according to claim 1, characterized in that the reference scan correction is in particular: The reference scanning data of the (x, ky) mixed domain comprises three rows of data, the corresponding positions of the 1 st row and the 3 rd row are averaged to obtain an interpolation row, and the interpolation row is multiplied by the conjugate of the 2 nd row of the reference scanning data of the (x, ky) mixed domain to obtain a plurality of rows of data; Obtaining phase difference information line data based on the complex line data fitting by using a first-order straight line fitting method; And removing the phase difference information line data from the forward degree skip in the image data in the (x, ky) mixed domain to obtain the first correction data.
6. The DWI-EPI image nyquist ghost correction method according to claim 5, characterized in that the first order straight line fitting method is a least square method.
7. The DWI-EPI image nyquist ghost correction method according to claim 1, characterized in that fourier transform is used for the x-direction of the first correction data in the (x, ky) mixed domain to obtain (Kx, ky) domain data.
8. The DWI-EPI image nyquist ghost correction method according to claim 1, wherein the no-reference scan correction method is an improved information entropy correction method, and the correction process includes: Generating a parameter range of a constant parameter b and a first-order linear parameter k of an information entropy correction method, and randomly generating a plurality of parameter combinations in the parameter range, wherein each parameter combination is used for one first-order linear phase error; Transforming the (Kx, ky) domain data into an (x, ky) mixed domain, and respectively carrying out phase error rejection operation on the mixed domain data by utilizing each first-order linear phase error to obtain multiple corrected data; Transforming a plurality of corrected data into an (x, y) image domain, calculating entropy values of corresponding images, and obtaining a parameter combination with the minimum entropy value as an initial parameter; And searching and obtaining an optimal parameter pair based on the initial parameters, and carrying out information entropy correction on the optimal parameter pair to obtain the second correction data.
9. The DWI-EPI image nyquist ghost correction method according to claim 8, wherein the optimal parameter pair is obtained by searching using a derivative-free method.

Description

DWI-EPI image Nyquist ghost correction method Technical Field The invention relates to the technical field of medical image processing, in particular to a DWI-EPI image Nyquist ghost correction method. Background In the world, nuclear magnetic resonance technology is rapidly developed, and is widely developed in a plurality of fields, such as low-field nuclear magnetism, and the low-field nuclear magnetic resonance technology is increasingly widely applied in the fields of food science, agriculture, petroleum energy, material science, textile chemical industry and the like. On the other hand, high field nuclear magnetic resonance is commonly used for human examination. Through years of technical development research and application practice research, magnetic resonance examination is almost applied to the whole body of a human body at present. Functional magnetic resonance imaging has been developed based on magnetic resonance to meet the needs of more examinations. High-field nuclear magnetic resonance (> = 1.5T) magnetic resonance Diffusion Weighted Imaging (DWI) is one of the functional magnetic resonance imaging, which is an integral part of the most advanced magnetic resonance imaging of modern times, indispensable in neuroimaging and oncology. DWI is a rapidly developing field of technology, and its application is increasing. It is currently a malignant lesion in acute ischemia, brain tumor, head and neck malignancy, breast cancer, and abdomen. EPI (Echo PLANAR IMAGING, planar Echo imaging) sequences are indispensable sequences in diffusion weighted imaging, which due to their extremely fast acquisition speed (which can be achieved in time millisecond acquisition to complete an image reconstruction)Spatial data), the DWI in clinic almost entirely adopts it as an acquisition modality. Although EPI has extremely fast imaging speeds, it is highly susceptible to magnetic field inhomogeneities to create a variety of artifacts, including mainly three types of artifacts, nyquist ghosting, chemical shift artifacts, and pattern distortion artifacts. The most common of these are Nyquist ghosts, because of e.g. gradient eddy currents,Field inhomogeneity, receive chain and gradient amplifier group delay, concomitant field, amplitude modulation andProblems such as spatial data displacement can cause Nyquist ghosting. This has an influence on the imaging accuracy. Disclosure of Invention The invention aims to overcome the defects of the prior art and provide a DWI-EPI image Nyquist ghost correction method with good correction effect and short correction time, especially for images with complicated phase information and difficult correction by using a traditional method. The aim of the invention can be achieved by the following technical scheme: a DWI-EPI image nyquist ghost correction method, comprising the steps of: acquiring magnetic resonance original data, dividing the magnetic resonance original data, extracting reference scanning data for reference scanning correction, filling the residual data in corresponding k space positions, and forming k space data in a (Kx, ky) domain; Transforming both the k-space data and the reference scan data in the (Kx, ky) domain into (x, ky) hybrid domain data; Performing reference scan correction on the (x, ky) mixed domain data to obtain first correction data; Transforming the first correction data in the (x, ky) mixed domain into (Kx, ky) domain data; Performing non-reference scanning correction method correction on the (Kx, ky) domain data to obtain second correction data; And performing zero padding processing on the second correction data to obtain third correction data, and performing two-dimensional inverse Fourier transformation on the third correction data to obtain an artifact-free image. Further, the segmentation of the magnetic resonance raw data is specifically: According to the sequence gradient information, the acquisition frequency and the system delay information, discarding the signal data acquired at noise points and read gradient climbing positions, only taking data points in a read gradient steady section in cyclic acquisition, and removing non-phase coding data, so that the data matrix size is consistent with the matrix size set during scanning. Further, the reference scan data is data with ky=0 in a data matrix acquired by scanning. Further, the negative gradient data in the k-space data is reversed left and right so that the k-space data are more uniform in direction when filling the corresponding k-space position. Further, inverse fourier transforming the k-space data and the reference scan data in the Kx direction, respectively, transforms each of the differently phase encoded signals into an (x, ky) hybrid domain. Further, the reference scan correction is specifically: The reference scanning data of the (x, ky) mixed domain comprises three rows of data, the corresponding positions of the 1 st row and the 3 rd row are averaged to obtain an interpolati