CN110322408A - Multicenter magnetic resonance image automated quality control method based on cloud platform - Google Patents
Multicenter magnetic resonance image automated quality control method based on cloud platform Download PDFInfo
- Publication number
- CN110322408A CN110322408A CN201910501603.2A CN201910501603A CN110322408A CN 110322408 A CN110322408 A CN 110322408A CN 201910501603 A CN201910501603 A CN 201910501603A CN 110322408 A CN110322408 A CN 110322408A
- Authority
- CN
- China
- Prior art keywords
- interest
- magnetic resonance
- image
- diffusion
- calculating
- 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.)
- Pending
Links
- 238000003908 quality control method Methods 0.000 title claims abstract description 23
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 claims abstract description 57
- 238000009792 diffusion process Methods 0.000 claims description 54
- 238000000034 method Methods 0.000 claims description 10
- 239000011159 matrix material Substances 0.000 claims description 9
- 210000004556 brain Anatomy 0.000 claims description 7
- 238000002598 diffusion tensor imaging Methods 0.000 claims description 6
- 239000006185 dispersion Substances 0.000 claims description 2
- 238000012800 visualization Methods 0.000 description 6
- 238000010586 diagram Methods 0.000 description 4
- 238000004458 analytical method Methods 0.000 description 2
- 238000002597 diffusion-weighted imaging Methods 0.000 description 2
- 239000004744 fabric Substances 0.000 description 2
- 238000003384 imaging method Methods 0.000 description 2
- 238000010606 normalization Methods 0.000 description 2
- 230000002159 abnormal effect Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000013499 data model Methods 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000011835 investigation Methods 0.000 description 1
- 238000002595 magnetic resonance imaging Methods 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000010415 tropism Effects 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
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
- 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
- G01R33/5602—Image enhancement or correction, e.g. subtraction or averaging techniques, e.g. improvement of signal-to-noise ratio and resolution by filtering or weighting based on different relaxation times within the sample, e.g. T1 weighting using an inversion pulse
-
- 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
- G01R33/565—Correction of image distortions, e.g. due to magnetic field inhomogeneities
-
- 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
- G01R33/58—Calibration of imaging systems, e.g. using test probes, Phantoms; Calibration objects or fiducial markers such as active or passive RF coils surrounding an MR active material
-
- G06T5/70—
-
- G06T5/80—
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/25—Determination of region of interest [ROI] or a volume of interest [VOI]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10072—Tomographic images
- G06T2207/10088—Magnetic resonance imaging [MRI]
Abstract
The multicenter magnetic resonance image automated quality control method based on cloud platform that the invention discloses a kind of includes the following steps: (1) by scanning obtained magnetic resonance image to water mould or subject；(2) magnetic resonance image that step (1) scanning obtains is transmitted in cloud platform；(3) quality-controlling parameters are chosen, the area-of-interest in magnetic resonance image is chosen, quality-controlling parameters is calculated in area-of-interest, obtain the calculated result of quality-controlling parameters.The present invention can quickly obtain the Quality Control results of polycentric magnetic resonance image automatically, and Quality Control results have more objectivity.The present invention can detect the stability that multicenter magnetic resonance scanner changes over time, and can assess magnetic resonance image quality, lay the foundation for the correctness of multicenter study result.
Description
Technical field
The present invention relates to magnetic resonance imaging and field of image processing more particularly to a kind of multicenter magnetic based on cloud platform are total
Vibration image automated quality control method.
Background technique
Currently, international hot spot brain planning item is just constantly pushing trans-regional, the united Brian Imaging of multicenter.Multicenter
The MR investigation of cooperation can shorten collection period with the diversity and scale of enlarged sample.From science, it is intended that
Experimental data between each center is comparable, the smaller the better by regional disparity and magnetic resonance software and hardware bring difference.Cause
This, the quality control of multicenter magnetic resonance image is significant to the accuracy and reliability for ensureing multicenter study result.
However, the magnetic resonance scanner that each centre scan uses, there is practical sex differernce in used magnetic resonance sequences etc., increase
The magnetic resonance image quality control of traditional single centre is generalized to polycentric difficulty and uncertainty.Researcher both domestic and external is
Spend great effort that the research controlled for magnetic resonance image quality is unfolded.Currently, the quality of magnetic resonance image, which controls, is mostly
Region of interest (Region of is observed by special magnetic resonance technician or manually delimited in water model data by visual observation
Interest, ROI) a small amount of parameter is roughly calculated, then assessed with technician personal experience.The mode of this quality control
A large amount of manpowers and time are consumed, and there is very strong subjectivity, the requirement to technician is very high, is not appropriate for being applied in mostly
In the magnetic resonance image quality control of the heart.Therefore, a set of unified standard is badly in need of in the quality control of polycentric magnetic resonance image
Automatic flow.
Summary of the invention
Due to lacking the standard and process of the magnetic resonance image quality accepted extensively control at present, the present invention is summarized and is referred to
Magnetic resonance image method of quality control in the past 10 years, devises the multicenter magnetic resonance image automated quality based on cloud platform
Control method.It can be used for carrying out detailed analysis to water mould magnetic resonance image, analyze each center magnetic resonance scanner and become at any time
The stability of change, the difference at more each center, it can also be used to detailed analysis be carried out to subject magnetic resonance image, assess image matter
Amount lays the foundation for the correctness of multicenter study result.
The technical solution adopted by the present invention to solve the technical problems is: a kind of multicenter magnetic resonance figure based on cloud platform
As automated quality control method, include the following steps:
(1) magnetic resonance image by being obtained to water mould or subject scanning；
(2) magnetic resonance image that step (1) scanning obtains is transmitted in cloud platform；
(3) quality-controlling parameters are chosen, the area-of-interest in magnetic resonance image are chosen, to quality in area-of-interest
Control parameter is calculated, and the calculated result of quality-controlling parameters is obtained.
Further, the quality-controlling parameters of selection are selected from the inhomogeneities (B0 inhomogeneous) of main field, letter
It makes an uproar than (Signal-to-Noise Ratio), Nyquist artifact (Nyquist Ghost ratio), distortion
(Distortion), Fractional anisotropy (Fractional Anisotropy, FA), apparent diffusion coefficient (Apparent
Diffusion Coefficient, ADC), normalizated correlation coefficient (Normalized Correlation, NC), main diffusion
The strong deviation in direction (Principal Direction, PD).
Further, for the calculating of main field inhomogeneities, the calculating of main field inhomogeneities, choosing are carried out using field figure
Taking the center of circle is water mould center, and radius is circular area-of-interest (the Region of of water mode radius N% (70≤N≤90)
Interest, ROI), the inhomogeneities of main field is calculated using formula (1),
Wherein, BOinhomoIndicate main field inhomogeneities, Max (ROI) indicates the maximum value in area-of-interest, Min
(ROI) minimum value in area-of-interest is indicated, Δ TE was indicated between two different echo times used in scanning process
Difference, f indicate centre frequency；
Calculating for signal-to-noise ratio schemes the calculating for carrying out signal-to-noise ratio using b0, by by two width b0 images of multiple scanning
Carry out pixel scale subtracts each other to obtain noise image, and the selections center of circle is water mould center, radius for water mode radius N% (70≤N≤
90) circle carries out the calculating of signal-to-noise ratio using formula (2) as area-of-interest,
Wherein, SNR indicates signal-to-noise ratio, SmeanFor the mean value of signal strength in signal pattern area-of-interest, σ is noise pattern
As the standard deviation in area-of-interest；
Calculating for Nyquist artefact level calculates Nyquist artifact using formula (3),
Wherein, GhostratioIndicate Nyquist artefact level, MmiddleThe expression center of circle is water mould center, and radius is water mould
The rounded interested area of radius N% (70≤N≤90), Mtop, MbottomThe length for indicating phase-encoding direction is water mode diameter
N% (70≤N≤90), width are the rectangle area-of-interest of water mode diameter K% (5≤K≤10), Mright, MleftIndicate frequency
The length of coding direction is water mode diameter N% (70≤N≤90), and width is that the rectangle sense of water mode diameter K% (5≤K≤10) is emerging
Interesting region；
For the calculating of distortion, formula (4) can be used to distort to calculate,
Wherein, Distortion indicates the stretching/compressing degree of image, diaPEIndicate image in the straight of phase-encoding direction
Diameter, diaFEIndicate image in the diameter in frequency coding direction；
Calculating for Fractional anisotropy, using formula (5) calculating section anisotropy,
Wherein, FA indicates Fractional anisotropy, and diffusion-weighted data are carried out with the fitting of diffusion tensor model, calculates
Tensor matrix out, and three characteristic value D of tensor matrix are found out to tensor diagonalization of matrix1, D2, D3, DavgIndicate D1, D2, D3Three
The mean value of person；The center of circle can be pre-defined at water mould center, radius is the area-of-interest of water mode radius N% (70≤N≤90), institute
The Fractional anisotropy asked is the mean value of all Fractional anisotropies in area-of-interest；
Calculating for apparent diffusion coefficient calculates the apparent diffusion coefficient of each Diffusion direction using formula (6)；
Sk=S0exp(-bADCk) (6)
Wherein, SkIndicate the diffusion weighted images of k-th of Diffusion direction, S0Indicate the image of no diffusion gradient, b table
Show the diffusion-weighted factor in scanning process, ADCkIndicate the apparent diffusion coefficient in k-th of direction；
Calculating for normalizated correlation coefficient is calculated using formula (7),
Wherein, NC indicates the value of normalizated correlation coefficient, AiFor the ith pixel value of image A, BiIt is i-th of image B
Pixel value, N are number of pixels total in piece image；
Calculating for the strong deviation of main dispersal direction (Principal Direction, PD), first with disperse
Amount imaging (Diffusion Tensor Imaging, DTI) model carries out models fitting to diffusion-weighted data, calculates automatically
The main dispersal direction of full brain, and visualized, so as to intuitively observe main dispersal direction in point of unit ball surface
Cloth.
Compared with the background technology, the present invention, it has the beneficial effect that
(1) by cloud platform, the quality control of real-time magnetic resonance image may be implemented, can quickly obtain as a result, being suitable for
Multicenter, the quality control of high-volume magnetic resonance image.
(2) control of magnetic resonance image quality is carried out using multiple determining parameters, Quality Control results more comprehensively, have more visitor
The property seen.
(3) it does not need special technician and carries out the control of magnetic resonance image quality, save manpower.
Detailed description of the invention
In order to illustrate the technical solution of the embodiments of the present invention more clearly, required use in being described below to embodiment
Attached drawing be briefly described.It is readily apparent that the accompanying drawings in the following description is only particular implementation as described in this application
Example, is not the limitation to protection scope of the present invention.For those of ordinary skill in the art, creative labor is not being paid
Under the premise of dynamic, some other embodiments and attached can be obtained with following examples according to the present invention and its attached drawing certainly
Figure.
Fig. 1 is the field figure and area-of-interest that use when calculating main field inhomogeneities by water model data；
Fig. 2 is the noise pattern and area-of-interest that use when calculating signal-to-noise ratio by water model data；
Fig. 3 is the b0 image and area-of-interest that use when calculating Nyquist artifact by water model data；
Fig. 4 is the main field inhomogeneities versus time curve calculated by water model data；
Fig. 5 is the signal-to-noise ratio versus time curve calculated by water model data；
Fig. 6 is the Nyquist artifact versus time curve calculated by water model data；
Fig. 7 is the distortion versus time curve calculated by water model data；
Fig. 8 is the Fractional anisotropy versus time curve calculated by water model data；
Fig. 9 is the radar map of the apparent diffusion coefficient composition in each disperse direction being calculated by water model data；
Figure 10 is the normalizated correlation coefficient curve graph between the continuous level being calculated by dynamic subject data without a head；
Figure 11 is by there is the dynamic normalizated correlation coefficient curve graph being tested between the continuous level that data are calculated of head；
It is main that Figure 12 is that the subject data obtained by the lesser Prisma machine scans of scanning bed vibration are calculated
Dispersal direction distribution map；
Figure 13 is the main expansion that the subject data obtained by the scanning bed biggish Trio machine scans of vibration are calculated
Dissipate directional spreding figure.
Specific embodiment
The present invention is based on cloud platforms to carry out specific parameter calculating to the magnetic resonance image that multicenter scans, thus right
Multicenter magnetic resonance image carries out automated quality control.In order to make those skilled in the art more fully understand the technical side in the application
Case, the present invention will be further explained below with reference to the accompanying drawings.With oneself of water mould and the diffusion-weighted imaging image of subject
The control of dynamicization quality is used as embodiment.But this is only a part of the embodiment of the application, instead of all the embodiments.It is based on
Specific embodiment described herein, others skilled in the art person without making creative work it is obtained its
His embodiment, should all fall within conception range of the invention.
The preferred embodiment of the present invention is described below with reference to attached drawing:
In general, the present invention disposes good corresponding cloud platform environment in magnetic resonance image quality control procedure, so that
Cloud platform can detect the magnetic resonance image of upload automatically.Corresponding quality-controlling parameters are chosen, when cloud platform detects upload
Magnetic resonance image when, choose magnetic resonance image in area-of-interest, in area-of-interest to quality-controlling parameters carry out
It calculates, obtains the calculated result of quality-controlling parameters, and calculated result is automatically saved.
Quality control is carried out by the diffusion weighted images that the water mould to same size scans, magnetic resonance can be measured and swept
Retouch the stability at any time of instrument.Select the inhomogeneities (B0 inhomogeneous) of main field, signal-to-noise ratio (Signal-to-
Noise Ratio), Nyquist artifact (Nyquist Ghost ratio) distorts (Distortion), Fractional anisotropy
(Fractional Anisotropy, FA), apparent diffusion coefficient (Apparent Diffusion Coefficient, ADC) are right
The diffusion weighted images of water mould carry out quality control.
Calculating for main field inhomogeneities carries out the calculating of main field inhomogeneities using field figure, chooses the center of circle and is
Water mould center, radius be water mode radius N% (70≤N≤90) circular area-of-interest (Region of Interest,
ROI), the schematic diagram of field figure and area-of-interest is as shown in Figure 1.The inhomogeneities of main field is calculated using formula (1),
Wherein, BOinhomoIndicate main field inhomogeneities, Max (ROI) indicates the maximum value in area-of-interest, Min
(ROI) minimum value in area-of-interest is indicated, Δ TE was indicated between two different echo times used in scanning process
Difference, f indicate centre frequency.
Calculating for signal-to-noise ratio is schemed to carry out signal-to-noise ratio using b0 for the influence for exempting disperse direction and diffusion gradient
Calculating.By subtracting each other to obtain noise image for the two width b0 images progress pixel scale of multiple scanning, the selection center of circle is water
Mould center, radius are the circle of water mode radius N% (70≤N≤90) as area-of-interest, noise image and region of interest
The schematic diagram in domain as shown in Fig. 2, carry out the calculating of signal-to-noise ratio using formula (2),
Wherein, SNR indicates signal-to-noise ratio, SmeanFor the mean value of signal strength in signal pattern area-of-interest, σ is noise pattern
As the standard deviation in area-of-interest.
Calculating for Nyquist artefact level is usually weighed with artifact relative to the percentage of master image gray-scale intensity
Nyquist artefact level is measured, since Nyquist artifact exists only in phase-encoding direction, therefore phase-encoding direction is taken
The mean value of the taken background pixel of the mean value of background pixel and frequency coding direction is subtracted each other, and gained difference is artifact strength.And with difference
The percentage of value and master image gray-scale intensity mean value calculates Nyquist artefact level.Calculate the b0 image used and 5 senses
As shown in figure 3, the phase-encoding direction of image is up and down direction, frequency coding direction is left and right directions for the division in interest region.
Nyquist artifact is calculated using formula (3),
Wherein, GhostratioIndicate Nyquist artefact level, MmiddleThe expression center of circle is water mould center, and radius is water mould
The rounded interested area of radius N% (70≤N≤90), Mtop, MbottomThe length for indicating phase-encoding direction is water mode diameter
N% (70≤N≤90), width are the rectangle area-of-interest of water mode diameter K% (5≤K≤10), Mright, MleftIndicate frequency
The length of coding direction is water mode diameter N% (70≤N≤90), and width is that the rectangle sense of water mode diameter K% (5≤K≤10) is emerging
Interesting region.
For the calculating of distortion, the inhomogeneities of main field can be more sensitively detected by distorting, can be used formula (4)
It distorts to calculate,
Wherein, Distortion indicates the stretching/compressing degree of image, diaPEIndicate image in the straight of phase-encoding direction
Diameter, diaFEIndicate image in the diameter in frequency coding direction.
Calculating for Fractional anisotropy, using formula (5) calculating section anisotropy,
Wherein, FA indicates Fractional anisotropy, and diffusion-weighted data are carried out with the fitting of diffusion tensor model, calculates
Tensor matrix out, and three characteristic value D of tensor matrix are found out to tensor diagonalization of matrix1, D2, D3, DavgIndicate D1, D2, D3Three
The mean value of person.The center of circle can be pre-defined at water mould center, radius is the area-of-interest of water mode radius N% (70≤N≤90), institute
The Fractional anisotropy asked is the mean value of all Fractional anisotropies in area-of-interest.
Calculating for apparent diffusion coefficient calculates the apparent diffusion coefficient of each Diffusion direction using formula (6),
Sk=S0exp(-bADCk) (6)
Wherein, SkIndicate the diffusion weighted images of k-th of Diffusion direction, S0Indicate the image of no diffusion gradient, b table
Show the diffusion-weighted factor in scanning process, ADCkIndicate the apparent diffusion coefficient in k-th of direction.
For using the disperse magnetic resonance weighted image of the same water mould of same sequence, phase can be calculated automatically by cloud platform
The parameter value answered.By the inhomogeneities of main field, signal-to-noise ratio, Nyquist artifact, distortion, the calculated result of Fractional anisotropy
Line chart is drawn by horizontal axis of the time, can show that parameters value changes with time rule, to judge magnetic resonance scanner
Stability at any time.Fig. 4 to Fig. 8 respectively indicates the master that some scanning center is calculated by the diffusion weighted images of water mould
The inhomogeneities in magnetic field, signal-to-noise ratio, Nyquist artifact, distortion, situation of change of the Fractional anisotropy in not same date, horizontal seat
Mark indicates the date, and ordinate indicates the value of parameters.Since used water mould is substantially uniformity, hydrone respectively spreads ladder
Degree direction apparent diffusion coefficient should be it is the same, can each Diffusion direction obtained by calculation apparent diffusion coefficient
Value observes the stability of diffusion gradient.As shown in figure 9, being calculated automatically for the diffusion weighted images of water mould by cloud platform
The apparent diffusion coefficient of each Diffusion direction out, and it is depicted as radar map, diffusion can intuitively be observed by the radar map
The stability of gradient.
When being scanned to subject, obtained diffusion weighted images are subjected to the influence of a variety of artifacts, such as the dynamic puppet of head
Shadow, directionality artifact etc..Head of the subject in repetition time (repetition time, TR) is dynamic to show as level on the image
Between gray-scale intensity it is abnormal.Normalizated correlation coefficient can be used between the continuous level of all Diffusion directions
(Normalized Correlation, NC) detects the relevant artifact of this intensity.Calculating for normalizated correlation coefficient,
Normalizated correlation coefficient is calculated using formula (7),
Wherein, NC indicates the value of normalizated correlation coefficient, AiFor the ith pixel value of image A, BiIt is i-th of image B
Pixel value, N are number of pixels total in piece image.The characteristics of normalizated correlation coefficient is presented in diffusion-weighted data be,
When b value is identical, the normalizated correlation coefficient curve approximation of different Diffusion gradient directions obeys same distribution, but works as and gray scale occur
When strength artifact, the normalizated correlation coefficient curve of different Diffusion gradient directions no longer obeys same distribution.It is total using same magnetic
Scanner, identical sequence and the parameter of shaking are to same subject progress twice sweep, and subject is dynamic without obvious head for the first time, second of subject
There is obvious head dynamic.The b0 figure of 6 different Diffusion directions of twice sweep image is calculated respectively automatically using cloud platform and is connected
NC value between subsequent layers face, and by result visualization.Continuous level is calculated using the scan data that first time is moved without obvious head
Between normalizated correlation coefficient curve graph it is as shown in Figure 10, be calculated continuously using the scan data for having obvious head dynamic for the second time
Normalizated correlation coefficient curve graph between level is as shown in figure 11, and abscissa indicates different level, and ordinate indicates normalization phase
Coefficient values.Comparison diagram 10 and Figure 11 are it is found that when without motion artifact, the normalization of the continuous level between different Diffusion directions
Related coefficient curve deviation is smaller, and each curve approximation obeys same distribution, and when head occur and moving artifact, different diffusion gradients
The normalizated correlation coefficient curve deviation of continuous level between direction is larger.
There may be directionality artifacts for diffusion-weighted imaging, are mainly shown as the main dispersal direction of measurement
The strong deviation of (Principal Direction, PD), this artifact are to be in weight in scanning process caused by the vibration of bed
It is more serious in 30kg crowd below.For detection direction artifact, we by cloud platform, first with dispersion tensor at
Picture (Diffusion Tensor Imaging, DTI) model carries out models fitting to diffusion-weighted data, calculates full brain automatically
Main dispersal direction, and visualized, so as to intuitively observe main dispersal direction in the distribution of unit ball surface.Just
In normal situation, main dispersal direction should be similar with brain structure in spherical surface distribution, that is, is divided into almost symmetrical point of left and right two
Cloth.It is not in this symmetry if there is directionality artifact.Siemens Prisma 3T and Siemens is used respectively
Trio 3T magnetic resonance scanner carries out the scanning of diffusion-weighted magnetic resonance image to subject, and Prisma gradient system is stronger, scanning
When, the with small vibration of bed is compared with Prisma, and Trio gradient system is weaker, and when scanning, the vibration of bed is big, it may appear that biggish side
Tropism artifact.For the diffusion weighted images of subject, by cloud platform, using diffusion tensor model to diffusion-weighted data
Models fitting is carried out, the main dispersal direction of full brain is calculated, spherical uniform is artificially divided into 3600 different directions, and calculate
The number of the main dispersal direction of all directions out, by result visualization.Use the main diffusion of Prisma scanning the data obtained
The visualization result in direction is as shown in figure 12, uses the visualization result such as figure of the main dispersal direction of Trio scanning the data obtained
Shown in 13, abscissa indicates that different directions, ordinate indicate the number of the main dispersal direction in all directions.Comparison diagram 12,
Figure 13 is it is found that when the non-directional artifact of disperse magnetic resonance weighted image, visualization result of main dispersal direction or so approximation
Symmetrically, when directionality artifact occurs in disperse magnetic resonance weighted image, half brain of left and right in the visualization result of main dispersal direction
Serious asymmetric situation to occur.
Claims (3)
1. a kind of multicenter magnetic resonance image automated quality control method based on cloud platform, which is characterized in that including walking as follows
It is rapid:
(1) magnetic resonance image by being obtained to water mould or subject scanning；
(2) magnetic resonance image that step (1) scanning obtains is transmitted in cloud platform；
(3) quality-controlling parameters are chosen, the area-of-interest in magnetic resonance image is chosen, quality is controlled in area-of-interest
Parameter is calculated, and the calculated result of quality-controlling parameters is obtained.
2. a kind of multicenter magnetic resonance image automated quality control method based on cloud platform according to claim 1,
It is characterized in that, the quality-controlling parameters of selection are selected from the inhomogeneities (B0 inhomogeneous) of main field, signal-to-noise ratio
(Signal-to-Noise Ratio), Nyquist artifact (Nyquist Ghost ratio), distortion (Distortion), portion
Divide anisotropy (Fractional Anisotropy, FA), apparent diffusion coefficient (Apparent Diffusion
Coefficient, ADC), normalizated correlation coefficient (Normalized Correlation, NC), main dispersal direction
The strong deviation of (Principal Direction, PD).
3. a kind of multicenter magnetic resonance image automated quality control method based on cloud platform according to claim 2,
It is characterized in that,
Calculating for main field inhomogeneities, the calculating of main field inhomogeneities is carried out using field figure, and the selection center of circle is water mould
Center, radius are the circular area-of-interest (Region of Interest, ROI) of water mode radius N% (70≤N≤90),
The inhomogeneities of main field is calculated using formula (1),
Wherein, BOinhomoIndicate main field inhomogeneities, Max (ROI) indicates the maximum value in area-of-interest, Min (ROI) table
Show the minimum value in area-of-interest, Δ TE indicates the difference between two different echo times used in scanning process, f table
Show centre frequency；
The calculating for carrying out signal-to-noise ratio is schemed in calculating for signal-to-noise ratio using b0, by carrying out two width b0 images of multiple scanning
Pixel scale subtracts each other to obtain noise image, and the selection center of circle is water mould center, and radius is water mode radius N%'s (70≤N≤90)
Circle is used as area-of-interest, and the calculating of signal-to-noise ratio is carried out using formula (2),
Wherein, SNR indicates signal-to-noise ratio, SmeanFor the mean value of signal strength in signal pattern area-of-interest, σ is noise image sense
Standard deviation in interest region；
Calculating for Nyquist artefact level calculates Nyquist artifact using formula (3),
Wherein, GhostratioIndicate Nyquist artefact level, MmiddleThe expression center of circle is water mould center, and radius is water mode radius
The rounded interested area of N% (70≤N≤90), Mtop, MbottomThe length for indicating phase-encoding direction is water mode diameter N%
(70≤N≤90), width are the rectangle area-of-interest of water mode diameter K% (5≤K≤10), Mright, MleftIndicate frequency coding
The length in direction is water mode diameter N% (70≤N≤90), and width is the rectangle region of interest of water mode diameter K% (5≤K≤10)
Domain；
For the calculating of distortion, formula (4) can be used to distort to calculate,
Wherein, Distortion indicates the stretching/compressing degree of image, diaPEIndicate image phase-encoding direction diameter,
diaFEIndicate image in the diameter in frequency coding direction；
Calculating for Fractional anisotropy, using formula (5) calculating section anisotropy,
Wherein, FA indicates Fractional anisotropy, and diffusion-weighted data are carried out with the fitting of diffusion tensor model, calculates and opens
Moment matrix, and three characteristic value D of tensor matrix are found out to tensor diagonalization of matrix1, D2, D3, DavgIndicate D1, D2, D3Three's
Mean value；The center of circle can be pre-defined at water mould center, radius is the area-of-interest of water mode radius N% (70≤N≤90), required
Fractional anisotropy is the mean value of all Fractional anisotropies in area-of-interest；
Calculating for apparent diffusion coefficient calculates the apparent diffusion coefficient of each Diffusion direction using formula (6)；
Sk=S0exp(-bADCk) (6)
Wherein, SkIndicate the diffusion weighted images of k-th of Diffusion direction, S0Indicate the image of no diffusion gradient, b expression is swept
The diffusion-weighted factor during retouching, ADCkIndicate the apparent diffusion coefficient in k-th of direction；
Calculating for normalizated correlation coefficient is calculated using formula (7),
Wherein, NC indicates the value of normalizated correlation coefficient, AiFor the ith pixel value of image A, BiFor the ith pixel of image B
Value, N is number of pixels total in piece image；
Calculating for the strong deviation of main dispersal direction (Principal Direction, PD), first with dispersion tensor at
Picture (Diffusion Tensor Imaging, DTI) model carries out models fitting to diffusion-weighted data, calculates full brain automatically
Main dispersal direction, and visualized, so as to intuitively observe main dispersal direction in the distribution of unit ball surface.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910501603.2A CN110322408A (en) | 2019-06-11 | 2019-06-11 | Multicenter magnetic resonance image automated quality control method based on cloud platform |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910501603.2A CN110322408A (en) | 2019-06-11 | 2019-06-11 | Multicenter magnetic resonance image automated quality control method based on cloud platform |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110322408A true CN110322408A (en) | 2019-10-11 |
Family
ID=68120841
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910501603.2A Pending CN110322408A (en) | 2019-06-11 | 2019-06-11 | Multicenter magnetic resonance image automated quality control method based on cloud platform |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110322408A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110853741A (en) * | 2019-11-07 | 2020-02-28 | 北京绪水互联科技有限公司 | Quality control method, device and system based on remote cooperation |
| CN113261940A (en) * | 2021-02-23 | 2021-08-17 | 上海市医疗器械检验研究院 | Method and device for detecting quality of magnetic resonance image |
| CN113985334A (en) * | 2021-11-08 | 2022-01-28 | 电子科技大学 | Method for evaluating signal-to-noise ratio of magnetic resonance scanning image |
Citations (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101322648A (en) * | 2008-07-29 | 2008-12-17 | 四川大学华西医院 | NMR imaging equipment stability and method for measuring image-forming index |
| CN101932599A (en) * | 2005-11-10 | 2010-12-29 | 阿波罗生命科学有限公司 | Molecule and chimeric molecule thereof |
| CN102018514A (en) * | 2010-12-30 | 2011-04-20 | 中国科学院深圳先进技术研究院 | Magnetic resonance diffusion tensor imaging method and system |
| CN102334991A (en) * | 2011-07-29 | 2012-02-01 | 泰山医学院 | Comprehensive test phantom for controlling quality of diffusion tensor magnetic resonance imaging |
| CN102336278A (en) * | 2010-05-03 | 2012-02-01 | 空中客车西班牙运营有限责任公司 | Automatic system for quality control and position correction of taped parts |
| CN104282021A (en) * | 2014-09-28 | 2015-01-14 | 深圳先进技术研究院 | Parameter error estimation method and device of magnetic resonance diffusion tensor imaging |
| CN105097856A (en) * | 2014-05-23 | 2015-11-25 | 全视科技有限公司 | Enhanced back side illuminated near infrared image sensor |
| CN105637536A (en) * | 2013-07-02 | 2016-06-01 | 外科信息科学股份有限公司 | Method and system for a brain image pipeline and brain image region location and shape prediction |
| CN105913458A (en) * | 2016-05-04 | 2016-08-31 | 浙江工业大学 | Alba fiber imaging method based on colony tracking |
| CN106780643A (en) * | 2016-11-21 | 2017-05-31 | 清华大学 | Magnetic resonance repeatedly excites diffusion imaging to move antidote |
| CN106997034A (en) * | 2017-04-25 | 2017-08-01 | 清华大学 | Based on the magnetic resonance diffusion imaging method that reconstruction is integrated by example of Gauss model |
| CN108324244A (en) * | 2018-01-03 | 2018-07-27 | 华东师范大学 | The construction method and system of automatic augmentation training sample for the diagnosis of AI+MRI Image-aideds |
| CN108366753A (en) * | 2015-10-07 | 2018-08-03 | 生物质子有限责任公司 | Selective sampling for assessing the structure space frequency with specified contrast mechanisms |
-
2019
- 2019-06-11 CN CN201910501603.2A patent/CN110322408A/en active Pending
Patent Citations (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101932599A (en) * | 2005-11-10 | 2010-12-29 | 阿波罗生命科学有限公司 | Molecule and chimeric molecule thereof |
| CN101322648A (en) * | 2008-07-29 | 2008-12-17 | 四川大学华西医院 | NMR imaging equipment stability and method for measuring image-forming index |
| CN102336278A (en) * | 2010-05-03 | 2012-02-01 | 空中客车西班牙运营有限责任公司 | Automatic system for quality control and position correction of taped parts |
| CN102018514A (en) * | 2010-12-30 | 2011-04-20 | 中国科学院深圳先进技术研究院 | Magnetic resonance diffusion tensor imaging method and system |
| CN102334991A (en) * | 2011-07-29 | 2012-02-01 | 泰山医学院 | Comprehensive test phantom for controlling quality of diffusion tensor magnetic resonance imaging |
| CN105637536A (en) * | 2013-07-02 | 2016-06-01 | 外科信息科学股份有限公司 | Method and system for a brain image pipeline and brain image region location and shape prediction |
| CN105097856A (en) * | 2014-05-23 | 2015-11-25 | 全视科技有限公司 | Enhanced back side illuminated near infrared image sensor |
| CN104282021A (en) * | 2014-09-28 | 2015-01-14 | 深圳先进技术研究院 | Parameter error estimation method and device of magnetic resonance diffusion tensor imaging |
| CN108366753A (en) * | 2015-10-07 | 2018-08-03 | 生物质子有限责任公司 | Selective sampling for assessing the structure space frequency with specified contrast mechanisms |
| CN105913458A (en) * | 2016-05-04 | 2016-08-31 | 浙江工业大学 | Alba fiber imaging method based on colony tracking |
| CN106780643A (en) * | 2016-11-21 | 2017-05-31 | 清华大学 | Magnetic resonance repeatedly excites diffusion imaging to move antidote |
| CN106997034A (en) * | 2017-04-25 | 2017-08-01 | 清华大学 | Based on the magnetic resonance diffusion imaging method that reconstruction is integrated by example of Gauss model |
| CN108324244A (en) * | 2018-01-03 | 2018-07-27 | 华东师范大学 | The construction method and system of automatic augmentation training sample for the diagnosis of AI+MRI Image-aideds |
Non-Patent Citations (2)
| Title |
|---|
| 李晨等: "多中心扩散 ＭＲＩ的质量控制", 《中国医学影像技术》 * |
| 连燕云: "神经导航中MRI图像脑肿瘤自动检测分割及DT-MRI研究", 《中国优秀博士学位论文全文数据库信息科技辑》 * |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110853741A (en) * | 2019-11-07 | 2020-02-28 | 北京绪水互联科技有限公司 | Quality control method, device and system based on remote cooperation |
| CN113261940A (en) * | 2021-02-23 | 2021-08-17 | 上海市医疗器械检验研究院 | Method and device for detecting quality of magnetic resonance image |
| CN113261940B (en) * | 2021-02-23 | 2024-03-15 | 上海市医疗器械检验研究院 | Method and device for detecting magnetic resonance image quality |
| CN113985334A (en) * | 2021-11-08 | 2022-01-28 | 电子科技大学 | Method for evaluating signal-to-noise ratio of magnetic resonance scanning image |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| JP5420206B2 (en) | Imaging parameter determination method, imaging parameter adjustment apparatus, computer-readable medium, and electronically readable data medium | |
| CN110322408A (en) | Multicenter magnetic resonance image automated quality control method based on cloud platform | |
| US6516210B1 (en) | Signal analysis for navigated magnetic resonance imaging | |
| JP5735793B2 (en) | System for quantitatively separating seed signals in MR imaging | |
| Ouyang et al. | High resolution magnetic resonance imaging of the calcaneus: age-related changes in trabecular structure and comparison with dual X-ray absorptiometry measurements | |
| US20170199258A1 (en) | Mr imaging using multi-echo k-space acquisition | |
| JPH0594540A (en) | Projecting method for forming two-dimensional image based on three-dimensional data | |
| DE112015003853T5 (en) | Parallel MR imaging with Nyquist ghost correction for EPI | |
| WO2010116124A1 (en) | Diffusion-weighted nuclear magnetic resonance imaging | |
| CN108693492A (en) | Magnetic resonance fingerprint for phase loop(PHC-MRF)System and method | |
| JP2018198682A (en) | Magnetic resonance imaging apparatus and magnetic resonance image processing method | |
| CN109959886A (en) | For determining the method, apparatus of the image quality information of MR imaging apparatus | |
| JP4927316B2 (en) | Magnetic resonance method | |
| JP4698231B2 (en) | Magnetic resonance diagnostic equipment | |
| CN114217255A (en) | Rapid liver multi-parameter quantitative imaging method | |
| CN105997074B (en) | A kind of magnetic resonance quantifies the more phase of echo approximating methods of susceptibility imaging | |
| US6100689A (en) | Method for quantifying ghost artifacts in MR images | |
| WO2006109550A1 (en) | Magnetic resonance imaging device and method | |
| JP6782681B2 (en) | Magnetic resonance imaging device, imaging parameter set generation arithmetic unit and imaging parameter set generation program | |
| Fujita et al. | Rigid real-time prospective motion-corrected three-dimensional multiparametric mapping of the human brain | |
| Murata et al. | Effect of hybrid of compressed sensing and parallel imaging on the quantitative values measured by 3D quantitative synthetic MRI: A phantom study | |
| JP2000279390A (en) | Magnetic resonance imaging device | |
| Callot et al. | Short‐scan‐time multi‐slice diffusion MRI of the mouse cervical spinal cord using echo planar imaging | |
| JP7321703B2 (en) | Image processing device and magnetic resonance imaging device | |
| US7078898B2 (en) | Methods and apparatus for magnetic resonance imaging |
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 | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20191011 |
|
| WD01 | Invention patent application deemed withdrawn after publication |