Middle and low field intensity magnetic sensitive weighted imaging optimization method and system

By employing high-pass filtering techniques involving parallel scanning and complex-domain phase separation, the problems of phase artifacts and low signal-to-noise ratio in low-to-medium field magnetic susceptibility weighted imaging are solved, achieving efficient image reconstruction and detail enhancement.

CN121304502AActive Publication Date: 2026-01-09安徽福晴医疗装备有限公司
View PDF 9 Cites 0 Cited by

Patent Information

Application Number
CN202511865804.2
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2025-12-11
Publication Date
2026-01-09
Estimated Expiration
2045-12-11

AI Technical Summary

Technical Problem

Low- and medium-field magnetic susceptibility weighted imaging suffers from problems such as numerous phase artifacts, long reconstruction time, high memory consumption, low signal-to-noise ratio, and local attenuation of magnetic susceptibility signals.

Method used

Parallel scanning techniques such as GRAPPA are employed, combined with dipole field projection and high-pass filtering based on complex domain phase separation. Through preliminary unwinding, background field separation, and high-frequency phase data extraction, phase artifacts are reduced and image details are improved.

Benefits of technology

It significantly reduces phase artifacts, improves image details, shortens reconstruction time, reduces memory consumption, and preserves local magnetic susceptibility signals.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN121304502A_ABST
    Figure CN121304502A_ABST
Patent Text Reader

Abstract

The invention discloses a medium and low field intensity magnetic sensitive weighted imaging optimization method and system, and relates to the field of magnetic sensitive weighted imaging, and the method comprises the steps: obtaining K space original data obtained by scanning a target object; reconstructing image domain data according to the K space original data, and extracting amplitude data and phase data of the image domain data; performing preliminary unwrapping processing on the phase data to obtain unwrapped phase data; performing background field separation on the unwrapping phase data to obtain a local phase diagram; performing high-pass filtering operation based on complex field phase separation on the local phase diagram and the amplitude data to obtain high-frequency phase data; and reconstructing a magnetic sensitive weighted image according to the high-frequency phase data and the amplitude data to obtain the magnetic sensitive weighted image of the target object. According to the method, the phase artifacts of the SWI image can be remarkably reduced, and the definition of image details is improved.
Need to check novelty before this filing date? Find Prior Art

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of magnetic susceptibility weighted imaging, and particularly relates to a method and system for optimizing low-field magnetic susceptibility weighted imaging. BACKGROUND

[0002] Susceptibility Weighted Imaging (SWI) is a technique that generates high-contrast images using the differences in magnetic susceptibility between tissues. It combines the magnitude information and phase data from complex images acquired by Magnetic Resonance Imaging (MRI), and enhances the sensitivity to local susceptibility changes through unique post-processing methods such as phase mask and minimum intensity projection. The differences in magnetic susceptibility of human tissues (such as deoxyhemoglobin in venous blood, iron deposition or calcification) can significantly affect the phase data, and SWI can clearly display small blood vessels, hemorrhage and calcification by highlighting these differences. Since the distribution of these substances is closely related to various pathological processes (such as cerebrovascular diseases and neurodegenerative diseases), SWI is widely used in the clinical diagnosis and evaluation of brain diseases.

[0003] The classic magnetic susceptibility weighted imaging and reconstruction usually adopts 3DGRE sequence to acquire data, reconstructs the image and obtains the original data containing magnitude and phase, generates a phase mask after unwrapping and high-pass filtering the original phase, and applies the mask to the magnitude multiple times; finally, the minimum intensity projection technique is used to project the processed three-dimensional data.

[0004] This method uses high-pass filtering for phase separation to remove low-frequency background interference, and must first perform a time-consuming phase unwrapping step to process the wrapped phase. This not only significantly prolongs the reconstruction time and increases the memory consumption, but also the unwrapping process is susceptible to noise and complex phase structures, which may cause error propagation and introduce artifacts. In addition, the zero padding operation is often required for frequency domain high-pass filtering, which may introduce strip artifacts and ringing effects. Finally, in order to ensure that the phase range is within [-π, π] for subsequent phase mask calculation, phase unwrapping processing is also required after filtering, which is a very lengthy process.

[0005] Moreover, since the signal-to-noise ratio increases with the increase of field strength, there is a natural disadvantage in the effect of low-field SWI. When removing the low-frequency background field, the setting of the cutoff frequency of high-pass filtering will indiscriminately filter out all spatial frequency components below the threshold. Although the magnetic field changes generated by local tissues are mainly high frequency, they still contain low frequency components related to the size of the object. When the filtering cutoff frequency is too high or the object size is large, these effective low-frequency signals that belong to the local field will be mistakenly removed, resulting in the attenuation or even disappearance of the local magnetic susceptibility signal, which manifests as a decrease in tissue contrast or the missed detection of small lesions. SUMMARY

[0006] To solve the technical problems in the background art, the present application provides a low-field susceptibility weighted imaging optimization method and system.

[0007] In a first aspect, the present application provides a low-field susceptibility weighted imaging optimization method and system, comprising: obtaining K-space raw data obtained by scanning a target object; reconstructing image domain data from the K-space raw data, and extracting amplitude data and phase data of the image domain data; performing preliminary unwrapping processing on the phase data to obtain unwrapped phase data; performing background field separation on the unwrapped phase data to obtain a local phase map; performing high-pass filtering operation on the local phase map and the amplitude data based on complex domain phase separation to obtain high-frequency phase data; reconstructing a susceptibility weighted image according to the high-frequency phase data and the amplitude data to obtain a susceptibility weighted image of the target object.

[0008] Preferably, the scanning mode is parallel scanning.

[0009] Preferably, the parallel scanning mode is GRAPPA parallel scanning mode.

[0010] Preferably, the background field separation on the unwrapped phase data to obtain a local phase map comprises: performing background field separation on the unwrapped phase data using a dipole field projection method to obtain a local phase map.

[0011] Preferably, the background field separation on the unwrapped phase data using a dipole field projection method to obtain a local phase map comprises: generating a binary mask according to the unwrapped phase data, and marking a local field region in the binary mask; performing background field removal on the local field region to obtain a local phase map.

[0012] Preferably, the high-pass filtering operation on the local phase map and the amplitude data based on complex domain phase separation to obtain high-frequency phase data comprises: performing complex reconstruction on the local phase map and the amplitude data to obtain complex data; extracting real part and imaginary part of the complex data; performing multiple 3D spatial smoothing operations on the real part and the imaginary part of the complex data respectively to obtain real part low-frequency data and imaginary part low-frequency data; combining the real part low-frequency data and the imaginary part low-frequency data to obtain low-frequency complex data; performing unitization conjugate operation on the low-frequency complex data; Multiply the low-frequency complex data, which has undergone normalization and conjugation, with the complex data to obtain the high-frequency data; High-frequency phase data is obtained by performing phase extraction on the high-frequency data.

[0013] Preferably, the magnetic susceptibility-weighted image of the target object is reconstructed based on high-frequency phase data and amplitude data, specifically including: Generate a phase mask based on high-frequency phase data; The phase mask is iterated and multiplied pixel by pixel with the amplitude data multiple times to obtain the original magnetic susceptibility weighted image; The original magnetic susceptibility weighted image is subjected to minimum density projection to obtain the magnetic susceptibility weighted image of the target object.

[0014] Secondly, this invention also proposes a low-to-medium field strength magnetically sensitive weighted imaging optimization system, comprising: The acquisition module is used to acquire the raw K-space data obtained by scanning the target object; The processing module is used to reconstruct image domain data from the original K-space data, extract amplitude and phase data from the image domain data; perform preliminary unwinding processing on the phase data to obtain unwound phase data; perform background field separation on the unwound phase data to obtain a local phase map; perform high-pass filtering operation based on complex domain phase separation on the local phase map and amplitude data to obtain high-frequency phase data; and reconstruct a magnetic susceptibility-weighted image based on the high-frequency phase data and amplitude data to obtain a magnetic susceptibility-weighted image of the target object.

[0015] Preferably, during the background field separation process, the dipole field projection method is used to separate the background field of the unwound phase data.

[0016] Preferably, a high-pass filtering operation based on complex domain phase separation is performed on the local phase map and amplitude data to obtain high-frequency phase data, specifically including: Complex data is obtained by complex reconstruction of the local phase map and amplitude data; Extract the real and imaginary parts of complex number data; Multiple 3D spatial smoothing operations are performed on the real and imaginary parts of the complex data to obtain low-frequency data of the real part and low-frequency data of the imaginary part. The real part low-frequency data and the imaginary part low-frequency data are combined to obtain low-frequency complex data; Perform a normalization-conjugation operation on low-frequency complex data; Multiply the low-frequency complex data, which has undergone normalization and conjugation, with the complex data to obtain the high-frequency data; High-frequency phase data is obtained by performing phase extraction on the high-frequency data.

[0017] Preferably, the scanning method is parallel scanning.

[0018] Preferably, the parallel scanning method is the GRAPPA parallel scanning method.

[0019] The proposed method and system for optimizing low-to-medium field strength magnetic susceptibility weighted imaging in this invention involves scanning the target object to acquire raw K-space data, reconstructing image domain data based on the raw K-space data, and extracting amplitude and phase data from the image domain data. Preliminary unwinding processing is performed on the phase data to obtain unwound phase data. Background field separation is performed on the unwound phase data to obtain a local phase map. High-pass filtering based on complex domain phase separation is applied to the local phase map and amplitude data to obtain high-frequency phase data. The magnetic susceptibility weighted image is reconstructed based on the high-frequency phase data and amplitude data to obtain a magnetic susceptibility weighted image of the target object. This method significantly reduces phase artifacts in SWI images and improves image details. Attached Figure Description

[0020] Figure 1 This is a flowchart illustrating the optimization method for medium-low field strength magnetic susceptibility weighted imaging in one embodiment of the present invention. Detailed Implementation

[0021] It should be noted that, unless otherwise specified, the embodiments and features described in the present invention can be combined with each other. The present invention will now be described in detail with reference to the accompanying drawings and embodiments.

[0022] Firstly, referring to Figure 1 The present invention proposes a method for optimizing magnetically sensitive imaging in low and medium field strength, comprising: Obtain the raw K-space data obtained by scanning the target object; Reconstruct image domain data from raw K-space data, and extract amplitude and phase data from the image domain data; The phase data is initially unwound to obtain unwound phase data; Background field separation is performed on the unwound phase data to obtain a local phase map; High-pass filtering is performed on the local phase map and amplitude data to obtain high-frequency phase data; The magnetic susceptibility weighted image of the target object is obtained by reconstructing the magnetic susceptibility weighted image based on the high-frequency phase data and amplitude data.

[0023] Since SWI imaging utilizes differences in magnetic susceptibility to highlight pathological findings, these differences are directly related to phase data. The processing of phase data directly affects image quality, influencing observation and diagnosis. Raw phase data includes entangled phases and low-frequency global background phases, all of which affect the representation of crucial local high-frequency phase data in susceptibility-weighted images.

[0024] It should be understood that SWI phase data is a superposition of local and background fields. While directly applying high-pass filtering to the amplitude and phase data in the image domain to remove low-frequency background can remove the low-frequency background field, it inevitably leads to signal loss of local field information. Especially since the background phase amplitude is enormous, any filtering or denoising operation performed before background removal will hinder the preservation of low-amplitude local field information.

[0025] This invention acquires raw K-space data by scanning the target object, reconstructs image domain data based on the raw K-space data, and extracts amplitude and phase data from the image domain data. First, the phase data undergoes preliminary unwinding processing, and then the obtained unwinding phase data is separated from the background field to obtain a local phase map. High-pass filtering is applied to the local phase map and amplitude data to obtain high-frequency phase data. Based on the high-frequency phase data and amplitude data, a magnetic susceptibility-weighted image is reconstructed to obtain a magnetic susceptibility-weighted image of the target object. This significantly reduces phase artifacts in SWI images and improves image details.

[0026] In this embodiment, the target object is scanned in parallel to speed up the scanning process.

[0027] In one specific embodiment, the parallel scanning method is the GRAPPA parallel scanning method.

[0028] In this embodiment, the raw K-space data is acquired by performing a parallel GRAPPA scan on the target object to speed up the scan and reduce motion artifacts.

[0029] It's important to understand that GRAPPA (GeneRalized Autocalibrating Partially Parallel Acquisitions) is a parallel imaging technique widely used in magnetic resonance imaging. It utilizes the spatial sensitivity information of multiple receiving coils to reconstruct missing K-space lines by acquiring partial data in K-space (below the Nyquist sampling rate) and using the correlation between signals acquired by adjacent coils. The core of this technique is to use a self-calibrating signal (usually acquiring a small number of additional ACS lines at the center of K-space) to directly calculate the reconstruction weights between coils from the actual scan data, without the need for separate pre-scans, thus significantly accelerating the scan speed or reducing motion artifacts.

[0030] K-space data is essentially a representation of frequency space, containing the raw data acquired during magnetic resonance imaging (MRI) scanning. In MRI, the magnetic field gradient excites the MRI signal via radio frequency pulses, forming a series of spatial domain data, including frequency and phase data, which are stored in K-space. By repeating the process of radio frequency pulses and gradient pulses in different gradient fields, a series of data points can be acquired in K-space. After K-space data acquisition, a Fourier transform is performed to convert K-space into image space, thus obtaining the final MRI image.

[0031] It should be understood that the local phase map in this embodiment is a phase map that only contains local phases.

[0032] In one specific embodiment, background field separation is performed on the unwound phase data to obtain a local phase map, specifically including: The background field of the unwound phase data was separated using the dipole field projection (PDF) method to obtain a local phase map.

[0033] It should be understood that PDF in this embodiment stands for Projection onto Dipole Fields, an algorithm specifically designed for phase unwinding and background field removal in quantitative magnetic susceptibility imaging research. It is based on the physical model that assumes the macroscopic background phase can be described by a dipole field model, where the dipole field is a typical manifestation of macroscopic magnetic field inhomogeneity in a uniform magnetic field. Through iterative optimization, the original phase data is projected onto the dipole field model, thereby separating the background phase (low frequency, global) and the local phase (high frequency, local).

[0034] In SWI (Short-Range Induction), the local magnetic field generated by the target tissue is superimposed on the background magnetic field, which originates from imperfect shimming of the device and differences in the magnetic susceptibility of the surrounding tissue. This interferes with the clear display of local details. When PDF (Picture-Based Image) simplifies the magnetic source to a magnetic dipole model, the magnetic field generated by the dipole source inside the local field is approximately orthogonal to the magnetic field generated by the dipole source outside the local field. Using the projection theorem of Hilbert space, PDF decomposes the background field within the local field and projects it onto the magnetic field space composed of the external dipole sources, thereby separating the target local field.

[0035] This embodiment employs the dipole field projection method to remove background from the phase data in strong magnetic susceptibility weighted imaging. This reduces signal loss in local fields, enhances details in magnetic susceptibility weighted images, effectively separates the target local field, and significantly improves the quality of magnetic susceptibility weighted images, especially in low and medium fields.

[0036] Specifically, the dipole field projection method is used to separate the background field of the unwound phase data to obtain a local phase map, which includes: A binary mask is generated based on the unwound phase data, and local field regions in the binary mask are marked. Background field removal is performed on local field regions to obtain local phase maps.

[0037] In this embodiment, the high-pass filtering operation is a high-pass filtering operation based on complex domain phase separation.

[0038] It's important to understand that high-pass filtering based on complex domain phase separation is a technique in MRI used to separate low-frequency information and achieve a high-pass filtering effect. The raw phase extracted during MRI imaging contains both low- and high-frequency information and suffers from phase entanglement. A common approach is to first untangle the entanglement and then use a window function for high-pass filtering. However, by separating the phase from the complex data, using multiple spatial smoothing filters to separate the low-frequency information, and then performing normalization and conjugation multiplication with the original complex data to extract the high-frequency phase information, thus removing the low-frequency information and achieving high-pass filtering, the untangling step can be eliminated.

[0039] In this embodiment, the high-pass filtering operation based on complex domain phase separation first combines the local phase map and amplitude data into complex data, then extracts the real and imaginary parts of the complex data and performs multiple spatial smoothing filters, and then combines the filtered real and imaginary parts to form low-frequency complex data; by constructing the normalized conjugate of the low-frequency complex signal and multiplying it with the original complex data, high-frequency phase information is extracted, thus achieving the effect of removing low-frequency background phase.

[0040] Therefore, in this embodiment, a high-pass filtering operation based on complex domain phase separation is performed on the local phase map and amplitude data to obtain high-frequency phase data, specifically including: Complex data is obtained by complex reconstruction of the local phase map and amplitude data; Extract the real and imaginary parts of complex number data; Multiple 3D spatial smoothing operations are performed on the real and imaginary parts of the complex data to obtain low-frequency data of the real part and low-frequency data of the imaginary part. The real part low-frequency data and the imaginary part low-frequency data are combined to obtain low-frequency complex data; Perform a normalization-conjugation operation on low-frequency complex data; Multiply the low-frequency complex data, which has undergone normalization and conjugation, with the complex data to obtain the high-frequency data; High-frequency phase data is obtained by performing phase extraction on the high-frequency data.

[0041] This embodiment employs a high-pass filtering operation based on complex domain phase separation, avoiding the 2π jump problem caused by phase map subtraction, and maintaining the output phase between [-π, π], which facilitates the calculation of the phase mask. Furthermore, the spatial smoothing filter uses boundary replication, which introduces fewer bar artifacts compared to the zero-padding operation of frequency domain filtering, reducing ringing effects. Finally, the method of directly processing the complex domain can naturally handle entangled phases, avoiding the necessary unwinding step before traditional high-pass filtering, reducing reconstruction time and memory consumption.

[0042] It is important to understand that the improved high-pass filtering operation in this embodiment, compared to the traditional homodyne filtering, can remove entanglement artifacts while preserving tissue details, and it boasts faster computation speed, ensuring the accuracy of low-frequency phase data separation. The dipole field projection method, being a high-order processing method, is applied to mid-to-low field SWI in this embodiment. By solving the inverse dipole problem, large-scale background phases are robustly separated from the total phase to improve image quality. Furthermore, applying the dipole field projection method before the high-pass filtering operation maximizes the preservation of fine vein structure details, avoids the loss of critical low-amplitude local phase information, and ensures that high-frequency information is not excessively removed, thus guaranteeing the complete extraction of high-frequency information and avoiding high-frequency information loss.

[0043] In another specific embodiment, background field separation is performed on the unwound phase data to obtain a local phase map, specifically including: The V-SHARP method was used to separate the background field of the unwound phase data to obtain a local phase map.

[0044] The V-SHARP (Variable-kernel Sophisticated Harmonic Artefact Reduction for Phase data) method in this embodiment is the variable kernel sphere mean method.

[0045] Specifically, the magnetic susceptibility-weighted image is reconstructed based on high-frequency phase and amplitude data to obtain the magnetic susceptibility-weighted image of the target object, including: Generate a phase mask based on high-frequency phase data; The phase mask is iterated and multiplied pixel by pixel with the amplitude data multiple times to obtain the original magnetic susceptibility weighted image; The original magnetic susceptibility weighted image is subjected to minimum density projection to obtain the magnetic susceptibility weighted image of the target object.

[0046] In specific implementation, a phase mask with a value in the interval [0, 1) is generated for the phase map in the high-frequency phase data with a range of [-π, π]. The phase mask is designed for the phase characteristics of a specific target tissue (such as venous blood containing paramagnetic substances). The phase mask is iteratively applied to the amplitude data, which can selectively and significantly attenuate the signal intensity of the target tissue region, thereby enhancing its contrast with the surrounding tissue in the amplitude data. The minimum density projection is used to project the amplitude maps of multiple adjacent layers in the original magnetic susceptibility weighted image along the direction perpendicular to the layer. The pixel value with the lowest signal intensity on the projection path is selected as the projection result, which further enhances the continuity of the target tissue structure (such as small veins) in three-dimensional space and highlights the low signal characteristics.

[0047] It is important to understand that Minimum Intensity Projection (MinIP) is a medical image post-processing method that generates a two-dimensional image by taking the minimum density value of each projection line of a three-dimensional image along a specific direction (such as coronal, sagittal, or any viewpoint).

[0048] In this embodiment, minimum density projection is used in susceptibility-weighted imaging (SWI) to highlight structures with extremely low signal intensity, thereby helping clinicians improve diagnostic efficiency.

[0049] Secondly, this invention also proposes a low-to-medium field strength magnetically sensitive weighted imaging optimization system, comprising: The acquisition module is used to acquire the raw K-space data obtained by scanning the target object; The processing module is used to reconstruct image domain data from the original K-space data, extract amplitude and phase data from the image domain data; perform preliminary unwinding processing on the phase data to obtain unwound phase data; perform background field separation on the unwound phase data to obtain a local phase map; perform high-pass filtering operation based on complex domain phase separation on the local phase map and amplitude data to obtain high-frequency phase data; and reconstruct a magnetic susceptibility-weighted image based on the high-frequency phase data and amplitude data to obtain a magnetic susceptibility-weighted image of the target object.

[0050] The scanning method in this embodiment is parallel scanning.

[0051] In one specific embodiment, the parallel scanning method is the GRAPPA parallel scanning method.

[0052] The processing module in this embodiment includes: The reconstruction unit is used to reconstruct image domain data based on the original K-space data, extract the amplitude and phase data of the image domain data, and perform preliminary unwinding processing on the phase data to obtain unwound phase data. The background field separation unit is used to separate the background field of the unwound phase data to obtain a local phase map. The high-pass filter unit is used to perform high-pass filtering on the local phase map and amplitude data to obtain high-frequency phase data; The magnetic susceptibility weighted imaging and reconstruction unit is used to reconstruct the magnetic susceptibility weighted image based on high-frequency phase data and amplitude data to obtain the magnetic susceptibility weighted image of the target object.

[0053] In one specific embodiment, the background field separation unit is a dipole field projection unit, which is used to perform background field separation on the unwound phase data using the dipole field projection method to obtain a local phase map.

[0054] In one specific embodiment, the background field separation unit is a VSHARP unit, and the VSHARP method is used to separate the background field of the unwound phase data to obtain a local phase map.

[0055] In one specific embodiment, the high-pass filtering operation is a high-pass filtering operation based on complex domain phase separation. That is, the high-pass filtering unit in this embodiment performs a high-pass filtering operation based on complex domain phase separation on the local phase map and amplitude data to obtain high-frequency phase data.

[0056] The present invention will now be described in conjunction with specific embodiments.

[0057] Example 1

[0058] This embodiment proposes a method for optimizing magnetically sensitive imaging in low to medium field strength, including: Obtain the raw K-space data obtained by scanning the target object; Reconstruct image domain data from raw K-space data, and extract amplitude and phase data from the image domain data; The phase data is initially unwound to obtain unwound phase data; The background field of the unwound phase data is separated by the dipole field projection method to obtain a local phase map. High-frequency phase data is obtained by performing a high-pass filter operation based on complex domain phase separation on the local phase map and amplitude data. The magnetic susceptibility weighted image of the target object is obtained by reconstructing the magnetic susceptibility weighted image based on the high-frequency phase data and amplitude data.

[0059] Example 2

[0060] This invention also proposes a low-to-medium field strength magnetic sensitivity weighted imaging optimization system, comprising: The acquisition module is used to acquire the raw K-space data obtained by performing a parallel GRAPPA scan on the target object; The processing module includes: The reconstruction unit is used to reconstruct image domain data based on the original K-space data, extract the amplitude and phase data of the image domain data, and perform preliminary unwinding processing on the phase data to obtain unwound phase data. The dipole field projection unit is used to separate the background field of the unwound phase data using the dipole field projection method to obtain a local phase map; The high-pass filter unit is used to perform high-pass filtering on the local phase map and amplitude data based on complex domain phase separation to obtain high-frequency phase data; The magnetic susceptibility weighted imaging and reconstruction unit is used to reconstruct the magnetic susceptibility weighted image based on high-frequency phase data and amplitude data to obtain the magnetic susceptibility weighted image of the target object.

[0061] In practice, the acquisition module acquires the raw K-space data obtained by performing GRAPPA parallel scanning on the target object and transmits the raw K-space data to the reconstruction unit. The reconstruction unit reconstructs image domain data based on the original K-space data, extracts the amplitude and phase data of the image domain data, performs preliminary unwrapping processing on the phase data to obtain unwrapped phase data, and passes the unwrapped phase data to the PDF unit and the amplitude data to the high-pass filtering unit. The PDF unit uses the dipole field projection method to separate the background field of the unwound phase data, obtain a local phase map, and then transmits the local phase map to the high-pass filter unit. The high-pass filtering unit performs a high-pass filtering operation based on complex domain phase separation on the local phase map and amplitude data to obtain high-frequency phase data, and then transmits the high-frequency phase data to the magnetic susceptibility weighted imaging and reconstruction unit. The magnetic susceptibility weighted imaging and reconstruction unit reconstructs the magnetic susceptibility weighted image based on high-frequency phase data and amplitude data to obtain the magnetic susceptibility weighted image of the target object.

[0062] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A method for optimizing magnetically sensitive weighted imaging in medium and low field strength, characterized in that, include: Obtain the raw K-space data obtained by scanning the target object; Reconstruct image domain data from raw K-space data, and extract amplitude and phase data from the image domain data; The phase data is initially unwound to obtain unwound phase data; Background field separation is performed on the unwound phase data to obtain a local phase map; High-frequency phase data is obtained by performing a high-pass filter operation based on complex domain phase separation on the local phase map and amplitude data. The magnetic susceptibility weighted image of the target object is obtained by reconstructing the magnetic susceptibility weighted image based on the high-frequency phase data and amplitude data.

2. The method for optimizing magnetically sensitive imaging in low and medium field strengths according to claim 1, characterized in that, The scanning method is parallel scanning.

3. The method for optimizing magnetically sensitive imaging in low and medium field strengths according to claim 2, characterized in that, The parallel scanning method is the GRAPPA parallel scanning method.

4. The method for optimizing magnetically sensitive imaging in medium and low field strength according to claim 1, characterized in that, Background field separation is performed on the unwound phase data to obtain a local phase map, specifically including: The background field of the unwound phase data is separated by the dipole field projection method to obtain a local phase map.

5. The method for optimizing magnetically sensitive imaging in medium and low field strengths according to claim 4, characterized in that, The background field of the unwound phase data is separated using the dipole field projection method to obtain a local phase map, specifically including: A binary mask is generated based on the unwound phase data, and local field regions in the binary mask are marked. Background field removal is performed on local field regions to obtain local phase maps.

6. The method for optimizing magnetically sensitive imaging in medium and low field strength according to claim 1, characterized in that, High-pass filtering based on complex domain phase separation is performed on the local phase map and amplitude data to obtain high-frequency phase data, specifically including: Complex data is obtained by complex reconstruction of the local phase map and amplitude data; Extract the real and imaginary parts of complex number data; Multiple 3D spatial smoothing operations are performed on the real and imaginary parts of the complex data to obtain low-frequency data of the real part and low-frequency data of the imaginary part. The real part low-frequency data and the imaginary part low-frequency data are combined to obtain low-frequency complex data; Perform a normalization-conjugation operation on low-frequency complex data; Multiply the low-frequency complex data, which has undergone normalization and conjugation, with the complex data to obtain the high-frequency data; High-frequency phase data is obtained by performing phase extraction on the high-frequency data.

7. The method for optimizing magnetically sensitive imaging in medium and low field strength according to claim 1, characterized in that, Based on high-frequency phase and amplitude data, a magnetic susceptibility-weighted image is reconstructed to obtain the magnetic susceptibility-weighted image of the target object, specifically including: Generate a phase mask based on high-frequency phase data; The phase mask is iterated and multiplied pixel by pixel with the amplitude data multiple times to obtain the original magnetic susceptibility weighted image; The original magnetic susceptibility weighted image is subjected to minimum density projection to obtain the magnetic susceptibility weighted image of the target object.

8. A medium-to-low field strength magnetically sensitive weighted imaging optimization system, characterized in that, include: The acquisition module is used to acquire the raw K-space data obtained by scanning the target object; The processing module is used to reconstruct image domain data from the original K-space data, extract amplitude and phase data from the image domain data; perform preliminary unwinding processing on the phase data to obtain unwound phase data; perform background field separation on the unwound phase data to obtain a local phase map; perform high-pass filtering operation based on complex domain phase separation on the local phase map and amplitude data to obtain high-frequency phase data; and reconstruct a magnetic susceptibility-weighted image based on the high-frequency phase data and amplitude data to obtain a magnetic susceptibility-weighted image of the target object.

9. The low-to-medium field strength magnetically sensitive weighted imaging optimization system according to claim 8, characterized in that, In the process of background field separation, the dipole field projection method is used to separate the background field of the unwound phase data.

10. The low-to-medium field strength magnetic sensitivity weighted imaging optimization system according to claim 8, characterized in that, High-pass filtering based on complex domain phase separation is performed on the local phase map and amplitude data to obtain high-frequency phase data, specifically including: Complex data is obtained by complex reconstruction of the local phase map and amplitude data; Extract the real and imaginary parts of complex number data; Multiple 3D spatial smoothing operations are performed on the real and imaginary parts of the complex data to obtain low-frequency data of the real part and low-frequency data of the imaginary part. The real part low-frequency data and the imaginary part low-frequency data are combined to obtain low-frequency complex data; Perform a normalization-conjugation operation on low-frequency complex data; Multiply the low-frequency complex data, which has undergone normalization and conjugation, with the complex data to obtain the high-frequency data; High-frequency phase data is obtained by performing phase extraction on the high-frequency data.

Citation Information

Patent Citations

  • Magnetic resonance imaging method and device

    CN104107045A

  • Phase processing method for parallel magnetic resonance imaging

    CN104749538A

  • Magnetic-resonance magnetic susceptibility weighted imaging post-processing method and system

    CN106960458A

  • Rapid scanning sequence for susceptibility weighted imaging and rapid susceptibility weighted imaging method

    CN108514415A

  • System and method for magnetic resonance imaging

    CN109477877A