CN106997034B - Based on the magnetic resonance diffusion imaging method rebuild using Gauss model as example integration - Google Patents
Based on the magnetic resonance diffusion imaging method rebuild using Gauss model as example integration Download PDFInfo
- Publication number
- CN106997034B CN106997034B CN201710277899.5A CN201710277899A CN106997034B CN 106997034 B CN106997034 B CN 106997034B CN 201710277899 A CN201710277899 A CN 201710277899A CN 106997034 B CN106997034 B CN 106997034B
- Authority
- CN
- China
- Prior art keywords
- diffusion
- gauss
- signal
- model
- collected
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
- G01R33/00—Arrangements or instruments for measuring magnetic variables
- G01R33/20—Arrangements or instruments for measuring magnetic variables involving magnetic resonance
- G01R33/44—Arrangements or instruments for measuring magnetic variables involving magnetic resonance using nuclear magnetic resonance [NMR]
- G01R33/48—NMR imaging systems
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
- G01R33/00—Arrangements or instruments for measuring magnetic variables
- G01R33/20—Arrangements or instruments for measuring magnetic variables involving magnetic resonance
- G01R33/44—Arrangements or instruments for measuring magnetic variables involving magnetic resonance using nuclear magnetic resonance [NMR]
- G01R33/48—NMR imaging systems
- G01R33/54—Signal processing systems, e.g. using pulse sequences ; Generation or control of pulse sequences; Operator console
- G01R33/56—Image enhancement or correction, e.g. subtraction or averaging techniques, e.g. improvement of signal-to-noise ratio and resolution
- G01R33/563—Image enhancement or correction, e.g. subtraction or averaging techniques, e.g. improvement of signal-to-noise ratio and resolution of moving material, e.g. flow contrast angiography
- G01R33/56341—Diffusion imaging
Landscapes
- Physics & Mathematics (AREA)
- High Energy & Nuclear Physics (AREA)
- Condensed Matter Physics & Semiconductors (AREA)
- General Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Vascular Medicine (AREA)
- General Health & Medical Sciences (AREA)
- Radiology & Medical Imaging (AREA)
- Engineering & Computer Science (AREA)
- Signal Processing (AREA)
- Magnetic Resonance Imaging Apparatus (AREA)
Abstract
The invention discloses a kind of based on the magnetic resonance diffusion imaging method rebuild using Gauss diffusion model as example integration.Wherein method includes: based on multilayer while predetermined sequence to be excited to carry out signal acquisition to measured target；Phase estimation is carried out to collected undersampled signal by parallel imaging technique；By the phase of estimation, collected undersampled signal, without diffusion-weighted reference picture, establish Gauss diffusion model；According to the directive undersampled signal of Gauss diffusion model integration institute, target equation is established；Target equation is solved using Nonlinear conjugate gradient algorithm iteration, obtains diffusion tensor parameter；Diffusion coefficient and diffusion weighted images is calculated according to diffusion tensor parameter.The high power that the embodiment of the present invention can be realized Diffusion Tensor Imaging as a result, accelerates acquisition, effectively reduces acquisition time, and can accurately estimate diffusion tensor parameter, obtains high s/n ratio, high-resolution diffusion image, meet clinical use demand.
Description
Technical field
The present invention relates to mr imaging technique fields, more particularly to one kind is based on using Gauss diffusion model as example integration
The magnetic resonance diffusion imaging method of reconstruction.
Background technique
Diffusion Tensor Imaging is capable of the micro-Brownian movement of Non-invasive detection body water, provides human fiber company
The functional information of the structural information and tissue that connect is a kind of important neuroimaging technology, all obtains in clinical and research
It is widely applied.Currently, Diffusion Tensor Imaging is the imaging technique for uniquely capableing of Non-invasive detection human nerve fibre bundle.
The contrast mechanisms of magnetic resonance diffusion imaging can use Gauss diffusion model, the diffusion in the Gauss diffusion model
Tensor parameter reflects the scattering nature of tissue.By diffusion tensor, diffusion system can directly be calculated, anisotropy expands
The important image index such as parameter is dissipated, for example, fractional anisotropy (Fractional Anisotropy, English abbreviation FA), average
Diffusion coefficient (Mean Diffusivity, English abbreviation MD), diffusion tensor characteristic direction.These image parameter indexs are applied
In the diagnosis that human brain, spinal cord, skeletal muscle tissue connect.
In the related technology, traditional Diffusion Tensor Imaging is compiled by applying the diffusion gradient in no less than 6 directions
Code, is calculated diffusion tensor by Gauss diffusion model.In order to obtain more accurate diffusion tensor, it usually needs more to expand
Coding gradient direction, such as 32 diffusion coding directions are dissipated, the signal acquisition time is considerably increased.In terms of resolution ratio, repeatedly
The acquisition of excitation plane echo sequence can reach the diffusion image for being higher than traditional single-shot technology resolution ratio, but repeatedly excitation
Acquisition also results in longer acquisition time.Therefore, although high-resolution, the acquisition of more diffusion gradients can be realized more accurate height
Mass diffusion imaging, but the required image scanning time is longer than the conventional method of current clinical use, limits neuroimaging
Diagnostic techniques is in clinical application and development.
Currently, the technology of accelerating magnetic resonance diffusion tensor imaging is broadly divided into two classes: 1) parallel imaging technique, parallel imaging
Technology has the signal of spatial sensitivities coding can be by lack sampling by solving aliasing algorithm by multi-channel coil acquisition
Signal restores to obtain complete image；However, parallel imaging technique receives the influence of channel coil count and signal-to-noise ratio at present,
Accelerate multiple usually at 2~3 times, and the signal-to-noise ratio of image is relatively low.(2) compressed sensing technology, compressed sensing technology utilize expansion
The sparsity for dissipating image data reduces acquisition number of signals, accelerates acquisition by applying sparsity constraints；However, depending on unduly
Compressed sensing technology frequently can lead to the problems such as image smoothing, loss in detail, and usually require random sampling, it is difficult to apply
In clinic.
Summary of the invention
The purpose of the present invention is intended to solve above-mentioned one of technical problem at least to a certain extent.
For this purpose, a kind of based on the magnetic resonance rebuild using Gauss diffusion model as example integration it is an object of the invention to propose
Diffusion imaging method.The high power that this method can be realized Diffusion Tensor Imaging accelerates acquisition, effectively reduces acquisition time,
And can accurately estimate diffusion tensor parameter, high s/n ratio, high-resolution diffusion image are obtained, meeting clinical use needs
It asks.
In order to achieve the above objectives, the embodiment of the present invention propose based on using Gauss diffusion model be example integration reconstruction magnetic
Resonance diffusion imaging method, comprising: excite predetermined sequence to carry out signal acquisition to measured target simultaneously based on multilayer；By parallel
Imaging technique carries out phase estimation to collected undersampled signal；Pass through the phase of estimation, the collected lack sampling letter
Number and without diffusion-weighted reference picture, establish Gauss diffusion model；All directions are integrated according to the Gauss diffusion model
Undersampled signal, establish the target equation for estimating diffusion tensor parameter；It is asked using Nonlinear conjugate gradient algorithm iteration
The target equation is solved, diffusion tensor parameter is obtained；Diffusion coefficient is calculated according to the diffusion tensor parameter and diffusion adds
Weight graph picture.
It is according to an embodiment of the present invention based on using Gauss diffusion model as example integration rebuild magnetic resonance diffusion imaging side
Method excites predetermined sequence to carry out signal acquisition to measured target simultaneously based on multilayer, and by parallel imaging technique to collecting
Undersampled signal carry out phase estimation, later, by the phase of estimation, collected undersampled signal and without diffusion plus
The reference picture of power establishes Gauss diffusion model, then, integrates the directive undersampled signal of institute according to Gauss diffusion model,
The target equation for estimating diffusion tensor parameter is established, and target equation is solved using Nonlinear conjugate gradient algorithm iteration,
Diffusion tensor parameter is obtained, finally, diffusion coefficient and diffusion weighted images is calculated according to diffusion tensor parameter, is at least had
Following advantages: 1) based on multilayer, the high power of excitation sequence accelerates to acquire simultaneously, can be realized the height of Diffusion Tensor Imaging
Accelerate acquisition again, effectively reduces acquisition time；2) under high power acceleration, it can accurately estimate diffusion tensor parameter, obtain height
Signal-to-noise ratio, high-resolution diffusion image, meet clinical use demand；3) it is suitable for the diffusion tensor of more dispersal directions acquisition
Imaging, can provide higher acceleration for the diffusion tensor imaging sequence of more direction；4) it is adopted by Gauss diffusion model foundation
Collect the relationship of signal and diffusion tensor parameter, can direct solution obtain diffusion tensor parameter, it is directly effective, reduce at multistep
Error caused by managing；5) there is good flexibility and robustness, be suitable for a variety of magnetic resonance sequences.
For the present invention only using Gauss model as example, the method actually proposed is a basic skills, can be extended
Integration to LDPC code is rebuild, such as diffusion kurtosis imaging (diffusion kurtosis imaging, DKI), gamma
Distributed model (Gamma distribution model) is truncated Gauss model (truncated Gaussian model), draws
It stretches exponential model (Stretched-Exponential model), more Gauss models (multi-Gaussian model), Q is empty
Between (q-space imaging) etc. is imaged, to meet the needs of different diffusion imaging technologies.
The additional aspect of the present invention and advantage will be set forth in part in the description, and will partially become from the following description
Obviously, or practice through the invention is recognized.
Detailed description of the invention
Above-mentioned and/or additional aspect and advantage of the invention will become from the following description of the accompanying drawings of embodiments
Obviously and it is readily appreciated that, in which:
Fig. 1 is according to an embodiment of the invention based on the magnetic resonance expansion rebuild using Gauss diffusion model as example integration
The flow chart of astigmatic image method；
Fig. 2 is the multilayer of the embodiment of the present invention while acquisition acceleration and the interior drop of layer being excited to adopt the exemplary diagram of acceleration；
Fig. 3 is the embodiment of the present invention based on the exemplary diagram for rebuilding process using Gauss diffusion model as the integration of example；
Fig. 4 is to integrate reconstruction technique by the Gauss diffusion model of Traditional parallel imaging and the embodiment of the present invention to estimate to obtain
Anisotropy coefficient comparison diagram.
Specific embodiment
The embodiment of the present invention is described below in detail, examples of the embodiments are shown in the accompanying drawings, wherein from beginning to end
Same or similar label indicates same or similar element or element with the same or similar functions.Below with reference to attached
The embodiment of figure description is exemplary, it is intended to is used to explain the present invention, and is not considered as limiting the invention.
Below with reference to the accompanying drawings being total to based on the magnetic rebuild using Gauss diffusion model as example integration for the embodiment of the present invention is described
Shake diffusion imaging method.
Fig. 1 is according to an embodiment of the invention based on the magnetic resonance expansion rebuild using Gauss diffusion model as example integration
The flow chart of astigmatic image method.
As shown in Figure 1, should be based on can be with using the magnetic resonance diffusion imaging method that Gauss diffusion model is example integration reconstruction
Include:
S110 excites predetermined sequence to carry out signal acquisition to measured target simultaneously based on multilayer.
It should be noted that the predetermined sequence can be it is preset according to actual needs, for example, the predetermined sequence can
To be echo planar imaging sequence etc..The embodiment of the present invention is not limited to particular sequence, can be flexibly applied to a variety of magnetic resonance and sweep
Retouch sequence.In embodiments of the present invention, it is illustrated by taking multiple excitation plane echo sequence as an example.
It is appreciated that excitation technique is that the novel magnetic resonance developed in recent years accelerates acquisition technique to multilayer simultaneously, technology is former
Reason be by and meanwhile excite multi-layer data and simultaneously carry out data acquisition and reduce sweep time.By multilayer while excitation sequence
The data of acquisition, actually in the data for selecting layer direction lack sampling, and the signal-to-noise ratio of data will not significantly decrease.For
This can excite predetermined sequence to carry out signal acquisition to measured target simultaneously based on multilayer in this step.
In order to further speed up acquisition, in one embodiment of the invention, by parallel imaging technique to collecting
Undersampled signal carry out phase estimation before, the magnetic resonance diffusion imaging method of the embodiment of the present invention may also include that based on more
Secondary excitation predetermined sequence carries out lack sampling acquisition in phase-encoding direction.
For example, by taking predetermined sequence is multiple excitation plane echo sequence as an example, based on multilayer while excitation plane
When echo sequence carries out signal acquisition to measured target, also lack sampling acquisition can be carried out in phase-encoding direction.Pass through this as a result,
Two kinds of technologies combine, and the acquisition of multiple excitation plane echo sequence can be realized high power acceleration, and multiple is accelerated to be equal to two methods
Drop adopts the product of multiple.For example, as in Fig. 2, by taking the echo planar imaging sequence acquisition that one excites three times as an example, if multilayer swashs simultaneously
It is 2 that hair technology, which accelerates multiple, and accelerating multiple in layer is 3, then the lack sampling multiple of final data is that excitation technique adds multilayer simultaneously
Accelerate the product of multiple in fast multiple and layer, i.e. the lack sampling multiple of final data is 6, it can be seen that, it is equivalent to and has reached 6 times
Acquisition accelerate.
In Diffusion Tensor Imaging, each direction can collect drop as shown in Figure 2 and adopt K space data, and
And the signal of multichannel can be obtained by the acquisition of multi-channel coil.Therefore, final acquisition data are multichannels, more diffused sheets
To lack sampling three-dimensional data.
It should be noted that conventional method in the prior art be by select layer and phase-encoding direction using parallel at
As technology carries out solution aliasing, but the data adopted are dropped for this high power, conventional method is unable to get accurate image and parameter.And
In order to solve this technical problem, in embodiments of the present invention, using the integration method for reconstructing based on Gauss diffusion model, to obtain
To accurate diffusion tensor parameter, to obtain the high-resolution diffusion image of high s/n ratio, specific implementation process can be subsequent
Embodiment is described in detail.
Further, in one embodiment of the invention, exciting predetermined sequence to measured target simultaneously based on multilayer
Before carrying out signal acquisition, the magnetic resonance diffusion imaging method of the embodiment of the present invention further include: when applying default to measured target
Between high gradient magnetic field, to be diffused gradient coding in each excitation, wherein the gradient direction of high gradient magnetic field is diffusion
Gradient coding direction.
Specifically, in magnetic resonance diffusion imaging, gradient coding can be diffused in each excitation.The coding can believed
Number acquisition before apply a period of time high gradient magnetic field, which is diffusion gradient coding direction.If expanding applying
Measured target produces physiological movement (for example, beating of cerebrospinal fluid) while dissipating coding gradient, and it will cause images to generate play
Strong phase change.In order to avoid final image generates serious aliasing artefacts, the embodiment of the present invention need to owe to adopt to collected
Sample signal carries out phase estimation, i.e. execution step S120.
S120 carries out phase estimation to collected undersampled signal by parallel imaging technique.
It in this step, can to the correction specific implementation of phase error are as follows: encoded by coil sensitivities parallel
Imaging technique directly rebuilds collected undersampled signal, obtains the phase for exciting correspondence image every time.
In an embodiment of the present invention, collected undersampled signal can be rebuild by following reconstruction formula:
Wherein, diFor need interpolation i-th of coil K space data；(m, n) is corresponding K space coordinate；(m',
It n' is) coordinate of the K space data in convolution kernel；i,i'∈[1,Nc]；NcFor port number；G is that the interpolation of parallel imaging is calculated
Son can be obtained by prescan；di'The K space data used for interpolation.By above-mentioned formula (1), all space K drops adopt position
Setting data can all be obtained by interpolation.
In order to obtain the image of high quality, further, in an embodiment of the present invention, in the phase estimated
Later, can also low-pass pictures filtering be carried out to the phase of the estimation.In this way, the phase of image is filtering it by low-pass pictures
Afterwards, it can satisfy the demand of phasing.This is because the phase error in diffusion image usually meets a smooth low order
Multinomial distribution.
As an example, the embodiment of the present invention can take Tukey window to carry out low-pass filtering in the space K, then extraction pair
The phase of image area is answered, provides phase estimation for subsequent Model Reconstruction algorithm.Wherein, Tukey window is rectangular window and more than two
The superposition of porthole, feature are the main secondary lobes of sampled signal than high, secondary lobe fast convergence.
S130 is established by the phase of estimation, collected undersampled signal and without diffusion-weighted reference picture
Gauss diffusion model.
Specifically, the signal of magnetic resonance diffusion imaging can be described by Gauss diffusion model, which can
It is established by the phase of estimation before, collected undersampled signal and without diffusion-weighted reference picture.
In an embodiment of the present invention, it in conjunction with the acquisition strategies in the embodiment of the present invention, can be established by following formula high
This diffusion model:
In above-mentioned formula (2), Ki,mIt can indicate m-th of diffusion coding direction, the K spacing wave positioned at the position i.Wherein, NpFor
Number of pixels；For without diffusion-weighted reference picture B0 in picture positionThe amplitude at place；D is to be joined by six diffusion tensors
The diffusion tensor matrices that number indicates, this six independent diffusion tensor parameters are respectively Dxx、Dyy、Dzz、Dxy、Dxz、Dyz, each
All there are six independent diffusion tensor parameters in picture point；BmThe B matrix that coding direction is spread for m-th, by diffusion gradient
Direction uniquely determines；The phase of coding direction is spread for m-th；Wherein, the diffusion-weighted of coding direction is spread m-th
Image can by by without diffusion-weighted reference picture B0 multiplied byAnd the exponential damping item as caused by diffusion gradientIt obtains；For K space encoding vector；For image space positions vector；For Fourier's operator, in multilayer
It excites in acquisition simultaneously, this is three-dimensional Fourier's operator, and multi-layer image is converted to three-dimensional K spacing wave.Pass through
Gauss model model, it is available in the spatial position K i, the data K of m dispersal directioni,m.In above-mentioned formula (2), to indicate letter
Clean, multi-channel data dimension is concealed, and actually f, p and K are multi-channel data.By above-mentioned formula (2), acquisition can establish
K spacing wave and diffusion tensor parameter between connection.
S140 is established according to the directive undersampled signal of Gauss diffusion model integration institute for estimating that diffusion tensor is joined
Several target equations.
In an embodiment of the present invention, the phase of the diffusion image of each Diffusion direction can pass through parallel imaging skill
Art is estimated to obtain, and no diffusion-weighted reference picture B0 can be obtained by fully sampled.Therefore, in above-mentioned formula (2) " phase " and
" no diffusion-weighted reference picture B0 " the two variables are known.Assuming that enough K spacing waves, above-mentioned formula can be obtained
(2) unique unknown quantity only has diffusion tensor parameter D in.It, can be by being based on Gauss in order to obtain diffusion tensor parameter D
The Diffusion Tensor Estimation of diffusion model obtains, and can integrate the directive undersampled signal of institute, by Gauss diffusion model to build
Found the target equation for estimating diffusion tensor parameter.Wherein, target equation can be established by following formula:
Wherein, first itemFor data fidelity term；Section 2 λ TV (D) be total variation sparsity about
Beam；To excite the down-sampled K spacing wave of high power collected with lack sampling in layer simultaneously by multilayer, including all expansions
Dissipate gradient direction；K (D) is the K spacing wave estimated by Gauss diffusion model (i.e. above-mentioned formula (2)), is uniquely become in estimation
Amount is diffusion tensor D；R be data selection operator, for select in K (D) withSignal with the consistent spatial position K is compared
Compared with；TV (D) is the total variation operator of 1 norm (i.e. L1-norm), can be acted in diffusion tensor parameter.Since diffusion tensor is joined
Number meets sparsity condition by total variation transformation, therefore the total variation sparsity constraints can effectively reduce image and make an uproar
Sound improves signal noise ratio (snr) of image, corresponding to the denoising operation in diffusion image post-processing algorithm.λ is control parameter, be can control complete
The contribution of variation sparsity constraints in the algorithm.
S150 solves target equation using Nonlinear conjugate gradient algorithm iteration, obtains diffusion tensor parameter.
After obtaining the target equation for estimating diffusion tensor parameter, Nonlinear conjugate gradient algorithm iteration can be used
The target equation is solved, to obtain diffusion tensor parameter.
For example, can estimate diffusion tensor parameter by process as shown in Figure 3 after establishing Gauss diffusion model.?
In the embodiment of the present invention, whole process can be divided into two parts: 1) Model Reconstruction pre-processes, and passes through parallel imaging and low-pass filtering
The image phase of each Diffusion direction is estimated, and is obtained by fully sampled without diffusion-weighted reference picture B0；2) sharp
Target equation is solved with Nonlinear conjugate gradient algorithm iteration, to obtain diffusion tensor parameter: the target equation based on foundation, repeatedly
In generation, solves diffusion tensor parameter D.The diffusion tensor parameter D estimated every time is by the available cost size of target equation, if generation
Valence has been restrained (if variation is less than 0.1%), then stops iteration and export the diffusion tensor parameter D currently estimated；Otherwise, it counts
The gradient for calculating target equation makes the smallest diffusion tensor parameter D of cost by the step length searching of setting along negative gradient direction,
And diffusion tensor parameter D is updated, it is repeated the above steps again by updated diffusion tensor parameter D, until the cost is received
It holds back.Wherein, D corresponding when convergence, diffusion tensor parameter as to be solved.
Diffusion coefficient and diffusion weighted images is calculated according to diffusion tensor parameter in S160.
After obtaining diffusion tensor parameter, by the diffusion weighted images of the available needs of diffusion tensor parameter,
The important image indexes such as fractional anisotropy, average diffusion coefficient and diffusion tensor characteristic direction parameter.
To sum up, during proposition is based on reconstruction is integrated as example using Gauss diffusion model, all diffusion coding staffs
To lack sampling data can all be used simultaneously in iterative approximation, the phase of different directions data in diffusion imaging is implicitly utilized
Guan Xing.Traditional diffusion tensor imaging needs to rebuild the data in each direction first, is then passed through by Gauss diffusion model more
Parametric regression algorithm obtains diffusion tensor.And the process of the independent image of reconstruction has been skipped in the embodiment of the present invention, pass through integration weight
Direct estimation diffusion tensor parameter is built, unknown quantity number is reduced, improves the accuracy of diffusion tensor parameter Estimation.For example,
Fig. 4 be by Traditional parallel imaging and the embodiment of the present invention Gauss diffusion model integrate reconstruction technique estimate it is each to different
Property coefficient comparison diagram.As shown in figure 4, with 8 times of acceleration acquisition data instance to 32 Diffusion directions, (wherein, multilayer swashs
It is 2 that hair, which accelerates multiple, in layer drop adopt accelerate multiple be 4), the embodiment of the present invention propose based on using Gauss diffusion model as example
The diffusion imaging method that integration is rebuild substantially increases the accuracy and signal-to-noise ratio of the diffusion tensor parameter Estimation under high power acceleration.
It is according to an embodiment of the present invention based on using Gauss diffusion model as example integration rebuild magnetic resonance diffusion imaging side
Method excites predetermined sequence to carry out signal acquisition to measured target simultaneously based on multilayer, and by parallel imaging technique to collecting
Undersampled signal carry out phase estimation, later, by the phase of estimation, collected undersampled signal and without diffusion plus
The reference picture of power establishes Gauss diffusion model, then, integrates the directive undersampled signal of institute according to Gauss diffusion model,
The target equation for estimating diffusion tensor parameter is established, and target equation is solved using Nonlinear conjugate gradient algorithm iteration,
Diffusion tensor parameter is obtained, finally, diffusion coefficient and diffusion weighted images is calculated according to diffusion tensor parameter, is at least had
Following advantages: 1) based on multilayer, the high power of excitation sequence accelerates to acquire simultaneously, can be realized the height of Diffusion Tensor Imaging
Accelerate acquisition again, effectively reduces acquisition time；2) under high power acceleration, it can accurately estimate diffusion tensor parameter, obtain height
Signal-to-noise ratio, high-resolution diffusion image, meet clinical use demand；3) it is suitable for the diffusion tensor of more dispersal directions acquisition
Imaging, can provide higher acceleration for the diffusion tensor imaging sequence of more direction；4) it is adopted by Gauss diffusion model foundation
Collect the relationship of signal and diffusion tensor parameter, can direct solution obtain diffusion tensor parameter, it is directly effective, reduce at multistep
Error caused by managing；5) there is good flexibility and robustness, be suitable for a variety of magnetic resonance sequences.
In the description of the present invention, it is to be understood that, term " first ", " second " are used for description purposes only, and cannot
It is interpreted as indication or suggestion relative importance or implicitly indicates the quantity of indicated technical characteristic.Define as a result, " the
One ", the feature of " second " can explicitly or implicitly include at least one of the features.In the description of the present invention, " multiple "
It is meant that at least two, such as two, three etc., unless otherwise specifically defined.
In the description of this specification, reference term " one embodiment ", " some embodiments ", " example ", " specifically show
The description of example " or " some examples " etc. means specific features, structure, material or spy described in conjunction with this embodiment or example
Point is included at least one embodiment or example of the invention.In the present specification, schematic expression of the above terms are not
It must be directed to identical embodiment or example.Moreover, particular features, structures, materials, or characteristics described can be in office
It can be combined in any suitable manner in one or more embodiment or examples.In addition, without conflicting with each other, the skill of this field
Art personnel can tie the feature of different embodiments or examples described in this specification and different embodiments or examples
It closes and combines.
Although the embodiments of the present invention has been shown and described above, it is to be understood that above-described embodiment is example
Property, it is not considered as limiting the invention, those skilled in the art within the scope of the invention can be to above-mentioned
Embodiment is changed, modifies, replacement and variant.
Claims (5)
1. a kind of based on the magnetic resonance diffusion imaging method rebuild using Gauss diffusion model as example integration, which is characterized in that packet
Include following steps:
Predetermined sequence is excited to carry out signal acquisition to measured target simultaneously based on multilayer；
Phase estimation is carried out to collected undersampled signal by parallel imaging technique；
By the phase of estimation, the collected undersampled signal and without diffusion-weighted reference picture, Gauss expansion is established
Dissipate model；Wherein, the Gauss diffusion model is established by following formula:
Wherein, NpFor number of pixels；For without diffusion-weighted reference picture B0 in picture positionThe amplitude at place；D is by six
The diffusion tensor matrices that a diffusion tensor parameter indicates；BmThe B matrix that coding direction is spread for m-th, by Diffusion direction
It determines；The phase of coding direction is spread for m-th；Wherein, m-th diffusion coding direction diffusion weighted images pass through by
It is described without diffusion-weighted reference picture B0 multiplied by describedAnd the exponential damping item as caused by diffusion gradient
It obtains；For Fourier's operator,For K space encoding vector；For image space positions vector；
According to the directive undersampled signal of Gauss diffusion model integration institute, the mesh for estimating diffusion tensor parameter is established
Mark equation；Wherein, the target equation is established by following formula:
Wherein, first itemFor data fidelity term；Section 2 λ TV (D) is total variation sparsity constraints item；
To excite the down-sampled K spacing wave of high power collected with lack sampling in layer, including all diffusion gradients simultaneously by multilayer
The data in direction；K (D) is the K spacing wave estimated by the Gauss diffusion model, and unique variable is to expand in the estimation
Dissipate tensor D；R be data selection operator, for select in K (D) withSignal with the consistent spatial position K is compared；TV
It (D) is the total variation operator of 1 norm；λ is control parameter；
The target equation is solved using Nonlinear conjugate gradient algorithm iteration, obtains diffusion tensor parameter；
Diffusion coefficient and diffusion weighted images is calculated according to the diffusion tensor parameter.
2. the method as described in claim 1, which is characterized in that it is described by parallel imaging technique to collected lack sampling
Before signal carries out phase estimation, the method also includes:
Lack sampling acquisition is carried out in phase-encoding direction based on multiple excitation predetermined sequence.
3. the method as described in claim 1, which is characterized in that excite predetermined sequence to tested mesh simultaneously based on multilayer described
Before mark carries out signal acquisition, the method also includes:
Apply the high gradient magnetic field of preset time to the measured target, to be diffused gradient coding in each excitation,
In, the gradient direction of the high gradient magnetic field is diffusion gradient coding direction.
4. the method as described in claim 1, which is characterized in that described to be believed by parallel imaging technique collected lack sampling
Number carry out phase estimation, comprising:
The parallel imaging technique encoded by coil sensitivities, rebuilds the collected undersampled signal, obtains every
The phase of secondary excitation signal.
5. method as claimed in claim 4, which is characterized in that by following reconstruction formula come to the collected lack sampling
Signal is rebuild:
Wherein, diFor need interpolation i-th of coil K space data；(m, n) is corresponding K space coordinate；(m', n') be
The coordinate of K space data in convolution kernel；i,i'∈[1,Nc]；NcFor port number；G is the interpolation operator of parallel imaging；di'For
The K space data that interpolation uses.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710277899.5A CN106997034B (en) | 2017-04-25 | 2017-04-25 | Based on the magnetic resonance diffusion imaging method rebuild using Gauss model as example integration |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710277899.5A CN106997034B (en) | 2017-04-25 | 2017-04-25 | Based on the magnetic resonance diffusion imaging method rebuild using Gauss model as example integration |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106997034A CN106997034A (en) | 2017-08-01 |
| CN106997034B true CN106997034B (en) | 2019-07-26 |
Family
ID=59434260
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710277899.5A Active CN106997034B (en) | 2017-04-25 | 2017-04-25 | Based on the magnetic resonance diffusion imaging method rebuild using Gauss model as example integration |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106997034B (en) |
Families Citing this family (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US10545211B2 (en) * | 2017-06-28 | 2020-01-28 | Synaptive Medical (Barbados) Inc. | Method of correcting gradient nonuniformity in gradient motion sensitive imaging applications |
| CN109342981B (en) * | 2018-09-20 | 2021-10-12 | 上海东软医疗科技有限公司 | Dynamic imaging method and device based on EPI sequence |
| CN110322408A (en) * | 2019-06-11 | 2019-10-11 | 浙江大学 | Multicenter magnetic resonance image automated quality control method based on cloud platform |
| CN113534032B (en) * | 2020-04-14 | 2023-01-31 | 上海联影医疗科技股份有限公司 | Magnetic resonance imaging method and system |
| CN112213674B (en) * | 2020-09-11 | 2023-03-21 | 上海东软医疗科技有限公司 | Magnetic resonance compressed sensing reconstruction method and device |
| CN112051531B (en) * | 2020-09-14 | 2022-10-28 | 首都医科大学附属北京天坛医院 | Multi-excitation navigation-free magnetic resonance diffusion imaging method and device |
| CN112581385B (en) * | 2020-11-30 | 2023-07-07 | 南方医科大学 | Diffusion kurtosis imaging tensor estimation method, medium and device based on multiple prior constraints |
| CN113313778B (en) * | 2021-05-13 | 2023-02-17 | 中国科学院深圳先进技术研究院 | Magnetic resonance image reconstruction method, computer device, and storage medium |
| CN114187374B (en) * | 2021-10-29 | 2024-06-04 | 厦门大学 | Multi-excitation plane echo diffusion weighted imaging method |
| CN117233676B (en) * | 2023-11-15 | 2024-03-26 | 之江实验室 | Echo time-dependent magnetic resonance diffusion imaging signal generation method and device |
| CN117890844B (en) * | 2024-03-15 | 2024-05-24 | 山东大学 | Magnetic resonance image reconstruction method based on optimized mask model |
Family Cites Families (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| DE4435104A1 (en) * | 1994-09-30 | 1996-04-04 | Siemens Ag | Pulse sequence with multi-layer excitation |
| DE19829850C2 (en) * | 1998-07-03 | 2000-06-15 | Siemens Ag | Reconstruction of a planar sectional image from magnetic resonance signals in inhomogeneous magnetic fields |
| CN103048632B (en) * | 2012-12-19 | 2015-11-25 | 北京大学 | A kind of quick fine angular resolution diffusion imaging method based on compressed sensing |
| CN103064046B (en) * | 2012-12-25 | 2015-04-15 | 深圳先进技术研究院 | Image processing method based on sparse sampling magnetic resonance imaging |
| CN104586394B (en) * | 2014-12-23 | 2017-04-12 | 中国科学院深圳先进技术研究院 | Method and system for removing magnetic resonance diffusion tensor imaging noise |
| CN104569880B (en) * | 2014-12-31 | 2017-04-05 | 中国科学院深圳先进技术研究院 | A kind of magnetic resonance fast imaging method and system |
| CN105548927B (en) * | 2015-08-06 | 2018-11-09 | 清华大学 | The magnetic resonance diffusion imaging method repeatedly excited excited simultaneously based on multilayer |
| CN105548928B (en) * | 2015-12-04 | 2018-04-24 | 厦门大学 | A kind of supper-fast MR imaging method of multilayer based on segmentation excitation space-time code |
| CN105467339A (en) * | 2015-12-31 | 2016-04-06 | 深圳先进技术研究院 | Quick multilayer magnetic resonance imaging method and device |
-
2017
- 2017-04-25 CN CN201710277899.5A patent/CN106997034B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN106997034A (en) | 2017-08-01 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106997034B (en) | Based on the magnetic resonance diffusion imaging method rebuild using Gauss model as example integration | |
| CN111513716B (en) | Method and system for magnetic resonance image reconstruction using an extended sensitivity model and a deep neural network | |
| JP6998218B2 (en) | MR imaging with motion detection | |
| CN103477238B (en) | Compressed sensing MR image reconstruction using constraint from prior acquisition | |
| WO2020019412A1 (en) | Magnetic resonance imaging method and apparatus, and device and storage medium | |
| US9746539B2 (en) | MR imaging with suppresion of flow artifacts | |
| US10203394B2 (en) | Metal resistant MR imaging | |
| CN106772167B (en) | Magnetic resonance imaging method employing and device | |
| CN105051563B (en) | Use parallel more section MR imagings of phase-modulation RF pulses | |
| CN106574954A (en) | Parallel mr imaging with nyquist ghost correction for epi | |
| US11415655B2 (en) | Reduced field-of-view perfusion imaging with high spatiotemporal resolution | |
| US10203387B2 (en) | MR imaging with enhanced susceptibility contrast | |
| US11002815B2 (en) | System and method for reducing artifacts in echo planar magnetic resonance imaging | |
| JP4698231B2 (en) | Magnetic resonance diagnostic equipment | |
| CN106796274B (en) | PROPELLER-MR imaging with artifact suppression | |
| US10871537B1 (en) | Systems and methods for background suppression in time-of-flight magnetic resonance angiography | |
| US9535148B2 (en) | Dynamic contrast enhanced magnetic resonance imaging with high spatial-temporal resolution | |
| CN109920017A (en) | The parallel MR imaging reconstructing method of the full variation Lp pseudonorm of joint from consistency based on feature vector | |
| EP2856196A2 (en) | System and method for magnetic resonance imaging using highly accelerated projection imaging | |
| US10818047B2 (en) | Iterative reconstruction of quantitative magnetic resonance images | |
| US10928475B2 (en) | Dynamic contrast enhanced magnetic resonance imaging with flow encoding | |
| JP4675936B2 (en) | Nuclear magnetic resonance imaging system | |
| WO2017167937A1 (en) | Dynamic mr imaging with increased temporal and spatial resolution | |
| Akçakaya et al. | Subject-specific convolutional neural networks for accelerated magnetic resonance imaging | |
| JP4906952B2 (en) | Nuclear magnetic resonance imaging system |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |