CN106960429B - A kind of CT image metal artifact bearing calibration and device - Google Patents
A kind of CT image metal artifact bearing calibration and device Download PDFInfo
- Publication number
- CN106960429B CN106960429B CN201710084349.1A CN201710084349A CN106960429B CN 106960429 B CN106960429 B CN 106960429B CN 201710084349 A CN201710084349 A CN 201710084349A CN 106960429 B CN106960429 B CN 106960429B
- Authority
- CN
- China
- Prior art keywords
- image
- projection
- metal
- preliminary corrections
- original
- 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
- 239000002184 metal Substances 0.000 title claims abstract description 79
- 238000012937 correction Methods 0.000 claims abstract description 98
- 238000000034 method Methods 0.000 claims abstract description 49
- 230000007423 decrease Effects 0.000 claims abstract description 26
- 238000001883 metal evaporation Methods 0.000 claims abstract description 20
- 238000012545 processing Methods 0.000 claims abstract description 16
- 230000011218 segmentation Effects 0.000 claims description 29
- 239000011159 matrix material Substances 0.000 claims description 7
- 230000002708 enhancing effect Effects 0.000 claims description 6
- 238000009499 grossing Methods 0.000 claims description 5
- 230000004927 fusion Effects 0.000 claims description 4
- 238000013461 design Methods 0.000 claims description 3
- 238000005457 optimization Methods 0.000 claims description 3
- 210000004872 soft tissue Anatomy 0.000 claims description 3
- 238000006467 substitution reaction Methods 0.000 claims description 3
- 230000017105 transposition Effects 0.000 claims description 3
- 238000012360 testing method Methods 0.000 claims description 2
- 230000002401 inhibitory effect Effects 0.000 claims 1
- 238000003706 image smoothing Methods 0.000 abstract description 4
- 238000001914 filtration Methods 0.000 description 10
- 230000000694 effects Effects 0.000 description 4
- 238000001228 spectrum Methods 0.000 description 3
- 230000002159 abnormal effect Effects 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 2
- 238000002591 computed tomography Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 210000000988 bone and bone Anatomy 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 238000003745 diagnosis Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 239000003814 drug Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000009659 non-destructive testing Methods 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 238000011084 recovery Methods 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 230000001629 suppression Effects 0.000 description 1
- 210000001519 tissue Anatomy 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/50—Image enhancement or restoration using two or more images, e.g. averaging or subtraction
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T3/00—Geometric image transformations in the plane of the image
- G06T3/08—Projecting images onto non-planar surfaces, e.g. geodetic screens
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/70—Denoising; Smoothing
-
- 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
- G06T2207/10081—Computed x-ray tomography [CT]
-
- 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
- G06T2207/20—Special algorithmic details
- G06T2207/20024—Filtering details
-
- 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
- G06T2207/20—Special algorithmic details
- G06T2207/20212—Image combination
- G06T2207/20221—Image fusion; Image merging
-
- 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
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30004—Biomedical image processing
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Apparatus For Radiation Diagnosis (AREA)
Abstract
The present invention provides a kind of CT image metal artifact bearing calibration and device, by under the constraint that priori projects, according to original projection and metal shadowing, Gaussian Energy decline processing is carried out to the metallic region iteration of original projection, obtain meeting tentatively not correcting projection and carry out backprojection reconstruction and obtaining preliminary corrections image for condition, it will be used to update prior image after preliminary corrections image smoothing and de-noising, prior image is updated by iteration, and each updated prior image is projected to obtain updated priori projection, under the constraint of priori projection in the updated, not correcting tentatively for the condition that meets is obtained to project and further obtain preliminary corrections image, until obtaining the preliminary corrections image for the condition that meets, it is merged with metal image, obtain correction of a final proof image.The present invention has effectively achieved the metal artifacts reduction of CT image, and avoids the rough problem in interpolation method interpolation boundary in the prior art and iterative method improvement time longer problem.
Description
Technical field
The present invention relates to field of image processings, more particularly, to a kind of CT image metal artifact bearing calibration and device.
Background technique
X ray computer tomography technology (Computed Tomography, CT) be a kind of very important medicine at
Image space formula and industrial nondestructive testing technology, are widely used in medical domain.When obtaining the CT image of patient, when detector issues
X-ray pass through human body there are when the position of metal object, the x-ray photon of low energy is significantly larger than by the ratio that metal object absorbs
The x-ray photon of high-energy.Therefore, the major part that detector receives is the x-ray photon of high energy, so that X-ray is averaged
Power spectrum can be mobile to high power spectrum direction, can generate serious metal artifacts so as to cause the CT image of backprojection reconstruction.These are pseudo-
Shadow has seriously affected doctor to the diagnostic analysis of abnormal structure.In order to obtain clear reliable CT image so that diagnosis makes
With need to be corrected to the metal artifacts in CT image.
The prior art generally uses interpolation method and iterative method to carry out the correction of metal artifacts.Interpolation method is utilized in projection domain
Projection value around metallic region carries out interpolation or according to certain similar or symmetry principle to metal puppet to metal artifacts region
Shadow zone domain carries out interpolation repairing in blocks.Such as: Kalender et al. is proposed with the projection value of metal shadowing's areas adjacent to metal puppet
Shadow zone domain carries out linear interpolation, and this method will lead to interpolation area and the boundary of neighbouring view field has uneven slip, so that
Image after correction has very serious secondary artifact.In addition, there is also this for polynomial interopolation, Wavelet Interpolation, normalization interpolation etc.
Class problem；Yang Cheng et al. proposes that they are partitioned into metal first according to the contaminated projection value of similarity criterion recovery
And artifact, then this two-part projection value is repaired according to similarity principle, this method is heavily dependent on
The effect of segmentation and the severity of artifact, the poor effect when picture structure or more complicated metal structure.Iterative method
It is then by view field's iterative backprojection CT image reconstruction except metal shadowing region.Though this method can be to a certain degree
Upper correction metal artifacts, but time complexity is very high, it is longer the time required to correction metal artifacts, it is not able to satisfy clinical demand.
Summary of the invention
The present invention provides a kind of CT image metal artifact bearing calibration and device to solve interpolation method interpolation in the prior art
The rough problem in boundary and iterative method improvement time longer problem.
According to an aspect of the present invention, a kind of CT image metal artifact bearing calibration is provided, comprising: step 1, utilize CT
The non-correcting CT image of original projection backprojection reconstruction in system carries out the first segmentation and to the non-correcting CT image respectively
Two segmentations, correspondence obtain metal image and prior image, are projected to obtain metal shadowing to the metal image；Step 2, right
The prior image is projected, and priori projection is obtained；Step 3, under the constraint of priori projection, according to the original throwing
Shadow and the metal shadowing carry out Gaussian Energy decline processing to the metallic region of the original projection, obtain preliminary corrections throwing
Shadow；Step 4, if the original projection and the difference of preliminary corrections projection are greater than the first preset threshold, with the preliminary school
Orthographic projection updates the original projection, executes step 3；If the original projection and the preliminary corrections projection difference be less than or
Equal to the first preset threshold, backprojection reconstruction preliminary corrections image is projected using the preliminary corrections；Step 5, if it is described preliminary
The root-mean-square error for correcting image and the prior image is greater than the second preset threshold, then with described preliminary after smoothed denoising
Prior image described in image update is corrected, step 2 is executed；If the root mean square of the preliminary corrections image and the prior image misses
Difference is less than or equal to the second preset threshold, and the preliminary corrections image is merged with the metal image, obtains final school
Positive image.
According to another aspect of the present invention, a kind of CT image metal artifact means for correcting is provided, comprising: backprojection reconstruction
Module, projection module, Gaussian Energy decline module, first judgment module and the second judgment module；Backprojection reconstruction module, is used for
Using the non-correcting CT image of original projection backprojection reconstruction in CT system, first point is carried out respectively to the non-correcting CT image
It cuts and obtains metal image and prior image with the second segmentation, correspondence, the metal image is projected to obtain metal shadowing；It throws
Shadow module obtains priori projection for carrying out the projection to the prior image；Gaussian Energy declines module, in institute
Under the constraint for stating priori projection, according to the original projection and the metal shadowing, to the metallic region of the original projection into
Row Gaussian Energy decline processing obtains not correcting projection tentatively；First judgment module, if for the original projection and it is described just
The difference that step does not correct projection is greater than the first preset threshold, does not correct the projection substitution original projection tentatively with described；If institute
It states original projection and the difference for not correcting projection tentatively is less than or equal to the first preset threshold, do not corrected tentatively using described
Project backprojection reconstruction preliminary corrections image；Second judgment module, if being used for the preliminary corrections image and the prior image
Root-mean-square error be greater than preset threshold, then with after smoothed denoising the preliminary corrections image replacement prior image；If institute
The root-mean-square error for stating preliminary corrections image and the prior image is less than or equal to the second preset threshold, by the preliminary corrections
Image is merged with the metal image, obtains correction of a final proof image.
A kind of CT image metal artifact bearing calibration proposed by the present invention and device, by under the constraint that priori projects,
According to original projection and metal shadowing, Gaussian Energy decline processing is carried out to the metallic region iteration of original projection, until original
Projection is less than or equal to the first preset threshold with the difference for not correcting projection tentatively, realizes the correction to original projection, and will
Original projection after correction, i.e. iteration are resulting tentatively not to correct projection, is projected to obtain preliminary corrections image, by preliminary school
For updating prior image after positive image smoothing denoising, so that the artifact and noise in updated prior image are all in certain journey
Weaken on degree, prior image is updated by iteration, and projected each updated prior image to obtain updated elder generation
Projection is tested, under the constraint of priori projection in the updated, obtain the first preset threshold condition of satisfaction does not correct projection tentatively simultaneously
Preliminary corrections image is further obtained, until the preliminary corrections image for meeting the second preset threshold condition is obtained, so that priori figure
Artifact as in is continuously removed, so that the artifact in preliminary corrections image is more effectively removed, by that will meet the
The preliminary corrections image of two preset threshold conditions is merged with metal image, obtains correction of a final proof image.The present invention effectively realizes
The metal artifacts reduction of CT image, and avoid the rough problem in interpolation method interpolation boundary in the prior art and iterative method is rectified
Longer problem of positive time.
Detailed description of the invention
Fig. 1 is the method flow diagram corrected according to the CT image metal artifact of the embodiment of the present invention；
Fig. 2 is the schematic device corrected according to the CT image metal artifact of the embodiment of the present invention.
Specific embodiment
With reference to the accompanying drawings and examples, specific embodiments of the present invention will be described in further detail.Implement below
Example is not intended to limit the scope of the invention for illustrating the present invention.
When obtaining the CT image of patient, when the X-ray that detector issues passes through human body there are when the position of metal object, meeting
The X that appearance X-ray beam hardening phenomenon, the i.e. x-ray photon of low energy are significantly larger than high-energy by the ratio that metal object absorbs
Ray photons, the major part that detector receives are the x-ray photons of high energy, so that the average power spectrum of X-ray can be to Gao Nengpu
Direction is mobile, can generate serious metal artifacts so as to cause the CT image of backprojection reconstruction.These artifacts have seriously affected doctor
The raw diagnostic analysis to abnormal structure.
As shown in Figure 1, the present invention provides a kind of CT image metal artifact bearing calibration, comprising: step 1, utilize CT system
In the non-correcting CT image of original projection backprojection reconstruction, the first segmentation and second point are carried out respectively to the non-correcting CT image
It cuts, correspondence obtains metal image and prior image, is projected to obtain metal shadowing to the metal image；Step 2, to described
Prior image is projected, and priori projection is obtained；Step 3, under the constraint of priori projection, according to the original projection and
The metal shadowing carries out Gaussian Energy decline processing to the metallic region of the original projection, obtains preliminary corrections projection；Step
Rapid 4, if the original projection and the difference of preliminary corrections projection are greater than the first preset threshold, projected with the preliminary corrections
The original projection is updated, step 3 is executed；If the original projection and the difference of preliminary corrections projection are less than or equal to the
One preset threshold projects backprojection reconstruction preliminary corrections image using the preliminary corrections；Step 5, if the preliminary corrections figure
The root-mean-square error of picture and the prior image is greater than the second preset threshold, then with the preliminary corrections figure after smoothed denoising
As updating the prior image, step 2 is executed；If the root-mean-square error of the preliminary corrections image and the prior image is less than
Or it is equal to the second preset threshold, the preliminary corrections image is merged with the metal image, obtains correction of a final proof image.
In the present embodiment, using the non-correcting CT image of original projection backprojection reconstruction in CT system, obtained non-school
Contain metal artifacts in positive CT image, needs further to be corrected.To carry out metal artifacts reduction, non-correcting CT image need to be carried out
First segmentation, obtains metal image, to be merged with subsequent and preliminary corrections image, obtains correction image；To not correcting
CT image carries out the second segmentation, obtains prior image, original for further iteration optimization to obtain priori projection after projecting
It projects, the corresponding preliminary corrections image of preliminary corrections projection of acquisition the first preset threshold of satisfaction, there are still secondary in the image
Artifact needs further to be corrected.By with the preliminary corrections image update prior image after smoothed denoising, and updating every time
Prior image corresponding priori projection constraint under, correct original projection, obtain preliminary corrections projection, and then obtain and meet the
The preliminary corrections image of two preset thresholds, which is merged with metal image can be obtained correction image, realize
The correction of CT image metal artifact.
A kind of CT image metal artifact bearing calibration proposed by the present invention, by under the constraint that priori projects, according to original
Begin projection and metal shadowing, Gaussian Energy decline processing carried out to the metallic region iteration of original projection, until original projection with
The difference for not correcting projection tentatively is less than or equal to the first preset threshold, realizes the correction to original projection, and will be after correction
Original projection, i.e., iteration is resulting does not tentatively correct projection, and progress backprojection reconstruction obtains preliminary corrections image, by preliminary school
For updating prior image after positive image smoothing denoising, so that the artifact and noise in updated prior image are all in certain journey
Weaken on degree, prior image is updated by iteration, and projected each updated prior image to obtain updated elder generation
Projection is tested, under the constraint of priori projection in the updated, obtain the first preset threshold condition of satisfaction does not correct projection tentatively simultaneously
Preliminary corrections image is further obtained, until the preliminary corrections image for meeting the second preset threshold condition is obtained, so that priori figure
Artifact as in is continuously removed, so that the artifact in preliminary corrections image is more effectively removed, by that will meet the
The preliminary corrections image of two preset threshold conditions is merged with metal image, obtains correction of a final proof image.The present invention effectively realizes
The metal artifacts reduction of CT image, and avoid the rough problem in interpolation method interpolation boundary in the prior art and iterative method is rectified
Longer problem of positive time.
Based on the above embodiment, optionally, in step 1, the first segmentation is carried out to carry out including the use of MeanShift algorithm
Filtering is enhanced and is split using single threshold split plot design；Carry out second segmentation be filtered including the use of Gaussian filter algorithm,
It is split using multi-threshold segmentation and is filled using the threshold value method of average.
Wherein, the first segmentation is carried out to be filtered enhancing including the use of MeanShift algorithm and utilize single threshold split plot design
It is split.
In the present embodiment, since metal object boundary and the bright artifact obscurity boundary of surrounding are unclear, metal object is not easy to be had
Effect is split.Therefore, certain enhancing processing need to be carried out to CT image, such as: gradient method, operator, high-pass filtering, mask matching
Method, statistics difference method and Mean Shift algorithm etc. make metal object and the other parts significant difference of CT image come, to have
Conducive to segmentation.In the present embodiment, specifically, enhancing processing is filtered to CT image using Mean Shift algorithm, so that
Metal object and the other parts significant difference of CT image come, and then extract metal image using single threshold segmentation, thus effectively
Ground Split has gone out metal object.
Define xiAnd zi, i=1,2 ..., n respectively indicate original image and filtering enhancing image, consider the space letter of image
Breath and grayscale information, any pixel point are represented by x=(xs,xr).Wherein, xsFor coordinate, xrFor gray value.So, it utilizes
Mean Shift algorithm enhances metallic region, can carry out by following Filtering Formula:
Wherein, w (xi) it is sampled point xiWeight, the sampled point x close from xiThere are biggish weight, i.e., sampling closer from x
Point, more effective to the statistical property around estimation x, vice versa.
Above formula Filtering Formula Kernel Function is defined as follows:
Wherein, C is normaliztion constant, hs, hrIt is wide for nucleus band.hsFor spatial bandwidth, value is bigger, and segmentation needs the time to get over
It is long；hrFor the gray scale bandwidth of pixel, value is bigger, and more image details will be ignored.In order to guarantee certain calculating
Speed and segmentation resolution ratio, hs, hrKing-sized value should not be chosen.
Detailed implementation detail using Mean Shift algorithm enhancing metallic region is as follows:
Step 1: initialization l=1, and make yi,l=xi；
Step 2: y is calculated according to above-mentioned Filtering Formulai,l+1=MS(yi,l)；
Step 3: step 2 is repeated until yi,lConverge to final result yi,L, then result is exported
Step 4: will converge to the one kind that is classified as of same a starting point, merges the very few class of pixel, obtains final result,
Simple threshold values method is taken to extract filtered metal image.
In the present embodiment, to using Mean Shift algorithm be filtered it is enhanced do not correct tentatively image carry out it is single
Metal image can be obtained in Threshold segmentation.
Wherein, the second segmentation is carried out to be filtered including the use of Gaussian filter algorithm, be split using multi-threshold segmentation
It is filled with using the threshold value method of average.
In the present embodiment, the prior image of better quality in order to obtain does not correct image tentatively to containing metal artifacts
It is filtered, filters out partial noise and streak artifact.Specifically, Gaussian filter algorithm is used, it will tentatively non-school
Each pixel and Gaussian kernel convolution sum of positive image are as output pixel value.Specifically size is used as 5 × 5 filtering
Core.Wherein, it is as follows to calculate mathematical expression form for filtering core weight matrix:
Wherein, δ=1.6；
In the present embodiment, to tentatively not correcting image and carried out multi-threshold segmentation after gaussian filtering.In the present embodiment
In, specifically, after multi-threshold segmentation obtain six air, soft tissue, normal tissue, bone tissue, metal, artifact parts.Pass through
With the mean CT-number of each part, each part for not correcting image tentatively after gaussian filtering obtains prior image.Wherein,
Metal and artifact sections use the mean CT-number of normal tissue to substitute.
Based on the above embodiment, optionally, in step 1, backprojection reconstruction is to utilize filter back-projection algorithm (Filter
Back-Projection, FBP) carry out backprojection reconstruction.
In the present embodiment, it is used for the metal image of fusion and the prior image for correcting artifact in order to obtain, needs
Utilize the non-correcting CT image of original projection backprojection reconstruction.It can be used instead using the non-correcting CT image of original projection backprojection reconstruction
Projection reconstruction algorithm, filter back-projection reconstruction algorithm or iterative backprojection algorithm for reconstructing etc..In the present embodiment, specifically, it adopts
With the non-correcting CT image of filter back-projection reconstruction algorithm backprojection reconstruction.The algorithm determines power as theoretical basis using center slice, counts
It calculates simply, quickly can obtain precision when the original projection of acquisition is affected by noise little by backprojection reconstruction image
Higher backprojection reconstruction image.
Based on the above embodiment, optionally, in step 3, Gaussian Energy decline includes gradient decline, ART process and non-
Negativity constraint.
In the present embodiment, under the constraint of priori projection, Gaussian Energy decline processing is carried out to original projection, specifically
Ground is declined by gradient so that the Gaussian Energy of metal artifacts declines, to achieve the effect that remove artifact, about by nonnegativity
Beam obtains best convergence point in gradient descent procedures, realizes that iteration carries out Gaussian Energy decline by ART process.
Based on the above embodiment, optionally, in steps of 5, smoothing denoising is using SL0 method.
In the present embodiment, since there are secondary artifacts in preliminary corrections image, in order in further iterative processing
Artifact is preferably corrected in journey, is used further to update prior image after need to carrying out preliminary corrections image smooth removal secondary artifact.
In the present embodiment, specifically, it is changed using Smoothde L0 (SL0) method to the secondary artifact in initial calibration image
Generation processing.
Based on the above embodiment, optionally, in steps of 5, according to by the product of the metal image and fusion factor, with
The correction of a final proof image is overlapped fusion.
In the present embodiment, specifically, image co-registration is carried out by following mathematic(al) representation:
Wherein, foutFor the correction of a final proof image, fcorTo project expiring for backprojection reconstruction after Gaussian Energy smooth correction
The preliminary corrections image of the second preset threshold of foot, fmetalFor the metal image, α is the image co-registration factor, T be it is described most
Metal CT value adjusting parameter in correction image eventually.
Based on the above embodiment, optionally, during Gaussian Energy decline, the mathematic(al) representation of gradient descent procedures are as follows:
Wherein,For projection of the kth time updated original projection after gradient declines,Forλ is that gradient declines step-length, ▽TTo seek transposition after the operation of single order forward difference, f is height
This function,δ=4, ▽ are that single order forward difference is asked to operate, and k is the number of iterations, xpIt is priori projection；
The mathematic(al) representation of ART process and nonnegativity restriction are as follows:
Wherein, xk+1For the updated original projection of kth time；For kth time updated original projection through gradient decline at
Projection after reason；H be one with projection identical dimensional matrix, and in original projection at metal shadowing's same position be 0,
It for being multiplied for 1, H and other matrixes is dot product operation, i.e., corresponding element multiplication in matrix battle array at his position；xoriFor original throwing
Shadow.
In the present embodiment, the mathematic(al) representation of Gaussian Energy decline process are as follows:
K=k+1；
Wherein, the mathematic(al) representation of gradient descent procedures are as follows:ART process and
The mathematic(al) representation of nonnegativity restriction are as follows:Wherein, tk+1For the power system in+1 iteration of kth
Number, k is the number of iterations,It is kth time updated original projection through gradient decline treated projection,For λ is under gradient
Step-length, ▽ dropTTo seek transposition after the operation of single order forward difference, f is Gaussian function,δ=4, ▽ are before seeking single order
To difference operation, xpIt is priori projection；xk+1For kth+1 time updated original projection；H is one and projection identical dimensional
Matrix, and be 0 at metal shadowing's same position in original projection, be with being multiplied for other matrixes for 1, H at other positions
Dot product operation, i.e., corresponding element is multiplied in matrix battle array；xoriFor original projection.
Based on the above embodiment, optionally, the mathematic(al) representation of smoothing denoising is carried out using SL0 method are as follows:
Wherein, y is the preliminary corrections image for being unsatisfactory for the second preset threshold, and x indicates that true picture, λ are between the two
Balance parameters, R (x) be regularization term.
Based on the above embodiment, optionally, the expression formula of regularization term R (x) are as follows:
Wherein, p and σ is dynamic optimization parameter, and p is dynamic Norm Control parameter, for adjusting the protection of soft tissue details
And noise suppression effect；σ is then used to control the degree of approximation of adjacent iteration twice.
Based on the above embodiment, optionally, in step 2, described to be projected as Siddon forward projection algorithm.
In the present embodiment, it is needed in the present embodiment, specifically for correcting the priori projection of original projection in order to obtain
Ground projects prior image using forward projection algorithm.The algorithm can also limit under Gaussian Energy while projection
The smoothness of drop.
As shown in Fig. 2, the present invention provides a kind of CT image metal artifact means for correcting, comprising: backprojection reconstruction module is thrown
Shadow module, Gaussian Energy decline module, first judgment module and the second judgment module；Backprojection reconstruction module, for utilizing CT
The non-correcting CT image of original projection backprojection reconstruction in system carries out the first segmentation and to the non-correcting CT image respectively
Two segmentations, correspondence obtain metal image and prior image, are projected to obtain metal shadowing to the metal image；Projective module
Block obtains priori projection for carrying out the projection to the prior image；Gaussian Energy declines module, in the elder generation
It tests under the constraint of projection, according to the original projection and the metal shadowing, the metallic region of the original projection is carried out high
This energy decline processing obtains not correcting projection tentatively；First judgment module, if for the original projection and it is described it is preliminary not
The difference for correcting projection is greater than the first preset threshold, does not correct the projection substitution original projection tentatively with described；If the original
Begin to project and be less than or equal to the first preset threshold with the difference for not correcting projection tentatively, does not correct projection tentatively using described
Backprojection reconstruction preliminary corrections image；Second judgment module, if for the equal of the preliminary corrections image and the prior image
Square error is greater than preset threshold, then replaces prior image with the preliminary corrections image after smoothed denoising；If described first
The root-mean-square error of step correction image and the prior image is less than or equal to the second preset threshold, by the preliminary corrections image
It is merged with the metal image, obtains correction of a final proof image.
A kind of CT image metal artifact means for correcting proposed by the present invention declines module and the first judgement by Gaussian Energy
Module, according to original projection and metal shadowing, carries out the metallic region iteration of original projection high under the constraint of priori projection
This energy decline processing, until original projection and the difference for not correcting projection tentatively are realized less than or equal to the first preset threshold
Correction to original projection, and by the original projection after correction, i.e., iteration is resulting does not correct projection, progress back projection tentatively
Reconstruction obtains preliminary corrections image, by the second judgment module, will be used to update priori figure after preliminary corrections image smoothing and de-noising
Picture passes through projection module, Gaussian Energy so that the artifact and noise in updated prior image all weaken to a certain extent
Decline module, first judgment module and the second judgment module, each updated prior image is projected after obtaining update
Priori projection, under the constraint of priori projection in the updated, obtain meet the first preset threshold condition tentatively do not correct throwing
Shadow simultaneously further obtains preliminary corrections image, so that the artifact in prior image and preliminary corrections image is continuously removed, directly
To the preliminary corrections image for meeting the second preset threshold condition is obtained, by the way that the preliminary corrections of the second preset threshold condition will be met
Image is merged with metal image, has obtained the CT image effectively corrected.The metal that the present invention has effectively achieved CT image is pseudo-
Shadow correction, and avoid the rough problem in interpolation method interpolation boundary in the prior art and the iterative method improvement time longer asks
Topic.
Finally, method of the invention is only preferable embodiment, it is not intended to limit the scope of the present invention.It is all
Within the spirit and principles in the present invention, any modification, equivalent replacement, improvement and so on should be included in protection of the invention
Within the scope of.
Claims (9)
1. a kind of CT image metal artifact bearing calibration characterized by comprising
Step 1, using the non-correcting CT image of original projection backprojection reconstruction in CT system, the non-correcting CT image is distinguished
The first segmentation and the second segmentation are carried out, correspondence obtains metal image and prior image, projected to obtain to the metal image
Metal shadowing；
Step 2, the prior image is projected, obtains priori projection；
Step 3, under the constraint of priori projection, according to the original projection and the metal shadowing, to the original throwing
The metallic region of shadow carries out Gaussian Energy decline processing, obtains preliminary corrections projection；
Step 4, if the original projection and the difference of preliminary corrections projection are greater than the first preset threshold, with the preliminary school
Orthographic projection updates the original projection, gos to step 3；If the difference that the original projection is projected with the preliminary corrections is less than
Or it is equal to the first preset threshold, backprojection reconstruction preliminary corrections image is projected using the preliminary corrections；
Step 5, if the root-mean-square error of the preliminary corrections image and the prior image is greater than the second preset threshold, with warp
Prior image described in the preliminary corrections image update after smoothing denoising, gos to step 2；If the preliminary corrections image and
The root-mean-square error of the prior image is less than or equal to the second preset threshold, by the preliminary corrections image and the metal figure
As being merged, correction of a final proof image is obtained.
2. the method according to claim 1, which is characterized in that in step 1, carry out first segmentation include be filtered enhancing and
It is split using single threshold split plot design；
The second segmentation is carried out to be filtered including the use of Gaussian filter algorithm, threshold value is split and utilized using multi-threshold segmentation
The method of average is filled.
3. the method according to claim 1, which is characterized in that in step 1, backprojection reconstruction is to utilize filter back-projection algorithm
Carry out backprojection reconstruction.
4. the method according to claim 1, which is characterized in that in step 3, Gaussian Energy decline includes gradient decline, ART mistake
Journey and nonnegativity restriction.
5. the method according to claim 1, which is characterized in that in steps of 5, smoothing denoising is using SL0 method.
6. method according to claim 4, which is characterized in that during Gaussian Energy decline, the mathematical table of gradient descent procedures
Up to formula are as follows:
Wherein,For projection of the kth time updated original projection after gradient declines,Fortk+1For the weight coefficient in+1 iteration of kth, tkFor the weight coefficient in kth time iteration, λ
Decline step-length for gradient,To seek transposition after the operation of single order forward difference, f is Gaussian function,δ=4,For
Single order forward difference is asked to operate, k is the number of iterations, xpIt is priori projection；
The mathematic(al) representation of ART process and nonnegativity restriction are as follows:
Wherein, xk+1For kth+1 time updated original projection；It is handled for the updated original projection of kth time through gradient decline
Projection afterwards；H be one with projection identical dimensional matrix, and in original projection at metal shadowing's same position be 0, other
It for being multiplied for 1, H and other matrixes is dot product operation, i.e., corresponding element multiplication in matrix battle array at position；xoriFor original projection.
7. method according to claim 5, which is characterized in that carry out the mathematic(al) representation of smoothing denoising using SL0 method are as follows:
Wherein, y is the preliminary corrections image for being unsatisfactory for the second preset threshold, and x indicates true picture, and λ is between the two flat
Weigh parameter, and R (x) is regularization term.
8. method according to claim 7, which is characterized in that the expression formula of regularization term R (x) are as follows:
Wherein, p and σ is dynamic optimization parameter, and p is dynamic Norm Control parameter, for adjusting the protection of soft tissue details and making an uproar
Sound inhibitory effect；σ is used to control the degree of approximation of adjacent iteration twice.
9. a kind of CT image metal artifact means for correcting characterized by comprising backprojection reconstruction module, projection module, Gauss
Energy declines module, first judgment module and the second judgment module；
Backprojection reconstruction module, for using the non-correcting CT image of original projection backprojection reconstruction in CT system, to it is described not
Correcting CT image carries out the first segmentation and the second segmentation respectively, and correspondence obtains metal image and prior image, to the metal figure
As being projected to obtain metal shadowing；
Projection module obtains priori projection for carrying out the projection to the prior image；
Gaussian Energy declines module, for being thrown under the constraint that the priori projects according to the original projection and the metal
Shadow carries out Gaussian Energy decline processing to the metallic region of the original projection, obtains not correcting projection tentatively；
First judgment module, if being greater than the first default threshold for the original projection and the difference for not correcting projection tentatively
Value does not correct the projection substitution original projection tentatively with described；If the original projection does not correct projection tentatively with described
Difference is less than or equal to the first preset threshold, does not correct projection backprojection reconstruction preliminary corrections image tentatively using described；
Second judgment module, if being greater than default threshold for the root-mean-square error of the preliminary corrections image and the prior image
Value then replaces prior image with the preliminary corrections image after smoothed denoising；If the preliminary corrections image and the elder generation
The root-mean-square error for testing image is less than or equal to the second preset threshold, and the preliminary corrections image and the metal image are carried out
Fusion, obtains correction of a final proof image.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710084349.1A CN106960429B (en) | 2017-02-16 | 2017-02-16 | A kind of CT image metal artifact bearing calibration and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710084349.1A CN106960429B (en) | 2017-02-16 | 2017-02-16 | A kind of CT image metal artifact bearing calibration and device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106960429A CN106960429A (en) | 2017-07-18 |
| CN106960429B true CN106960429B (en) | 2019-08-27 |
Family
ID=59481639
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710084349.1A Active CN106960429B (en) | 2017-02-16 | 2017-02-16 | A kind of CT image metal artifact bearing calibration and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106960429B (en) |
Families Citing this family (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107492132B (en) * | 2017-08-31 | 2020-12-15 | 南方医科大学 | Image artifact removing method of digital tomography system |
| CN108074223B (en) * | 2017-12-28 | 2021-09-28 | 中国矿业大学(北京) | Automatic extraction method of crack network in coal rock sequence CT (computed tomography) image |
| CN109697691B (en) * | 2018-12-27 | 2022-11-25 | 重庆大学 | Dual-regularization-term-optimized finite-angle projection reconstruction method based on L0 norm and singular value threshold decomposition |
| CN110473269B (en) * | 2019-08-08 | 2023-05-26 | 上海联影医疗科技股份有限公司 | Image reconstruction method, system, equipment and storage medium |
| JP7317651B2 (en) * | 2019-09-24 | 2023-07-31 | 富士フイルムヘルスケア株式会社 | MEDICAL IMAGE PROCESSING APPARATUS AND MEDICAL IMAGE PROCESSING METHOD |
| CN110796620B (en) * | 2019-10-29 | 2022-05-17 | 广州华端科技有限公司 | Interlayer artifact suppression method and device for breast tomographic reconstruction image |
| CN111127477B (en) * | 2019-12-12 | 2023-10-20 | 明峰医疗系统股份有限公司 | CT polychromatic spectrum voxel simulation method |
| CN111815521B (en) * | 2020-05-27 | 2024-02-13 | 南京国科精准医学科技有限公司 | Cone beam CT metal artifact correction algorithm based on priori image |
| CN113256542B (en) * | 2021-06-22 | 2021-10-01 | 明峰医疗系统股份有限公司 | Noise suppression method, system and medium for CT scanner |
| CN114332270B (en) * | 2021-12-02 | 2023-03-31 | 赛诺威盛科技(北京)股份有限公司 | CT image metal artifact removing method and device for minimally invasive interventional surgery |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102567958A (en) * | 2010-12-31 | 2012-07-11 | 东软飞利浦医疗设备系统有限责任公司 | Image postprocessing method for removing metal artifact from computed tomography (CT) image |
| CN103020928A (en) * | 2012-11-21 | 2013-04-03 | 深圳先进技术研究院 | Metal artifact correcting method of cone-beam CT (computed tomography) system |
| CN103679642A (en) * | 2012-09-26 | 2014-03-26 | 上海联影医疗科技有限公司 | Computerized tomography (CT) image metal artifact correction method, device and computerized tomography (CT) apparatus |
| CN103745440A (en) * | 2014-01-08 | 2014-04-23 | 中国科学院苏州生物医学工程技术研究所 | Metal artifact correction method for CT (computerized tomography) systems |
| CN104821003A (en) * | 2015-04-13 | 2015-08-05 | 中国科学院苏州生物医学工程技术研究所 | CT image reconstruction method |
| CN104992409A (en) * | 2014-09-30 | 2015-10-21 | 中国科学院苏州生物医学工程技术研究所 | CT image metal artifact correction method |
| CN106296615A (en) * | 2016-08-16 | 2017-01-04 | 广州华端科技有限公司 | CT image corrects the method and system of metal artifacts |
-
2017
- 2017-02-16 CN CN201710084349.1A patent/CN106960429B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102567958A (en) * | 2010-12-31 | 2012-07-11 | 东软飞利浦医疗设备系统有限责任公司 | Image postprocessing method for removing metal artifact from computed tomography (CT) image |
| CN103679642A (en) * | 2012-09-26 | 2014-03-26 | 上海联影医疗科技有限公司 | Computerized tomography (CT) image metal artifact correction method, device and computerized tomography (CT) apparatus |
| CN103020928A (en) * | 2012-11-21 | 2013-04-03 | 深圳先进技术研究院 | Metal artifact correcting method of cone-beam CT (computed tomography) system |
| CN103745440A (en) * | 2014-01-08 | 2014-04-23 | 中国科学院苏州生物医学工程技术研究所 | Metal artifact correction method for CT (computerized tomography) systems |
| CN104992409A (en) * | 2014-09-30 | 2015-10-21 | 中国科学院苏州生物医学工程技术研究所 | CT image metal artifact correction method |
| CN104821003A (en) * | 2015-04-13 | 2015-08-05 | 中国科学院苏州生物医学工程技术研究所 | CT image reconstruction method |
| CN106296615A (en) * | 2016-08-16 | 2017-01-04 | 广州华端科技有限公司 | CT image corrects the method and system of metal artifacts |
Non-Patent Citations (3)
| Title |
|---|
| A prior-based metal artifact reduction algorithm for x-ray CT;Ming Li 等;《Journal of X-Ray Science and Technology》;20150430;第23卷(第2期);第229-241页 * |
| CT金属伪影校正研究;李铭;《中国博士学位论文全文数据库 信息科技辑》;20190915(第9期);第I138-63页 * |
| Gaussian diffusion sinogram inpainting for X‑ray CT metal artifact reduction;Chengtao Peng 等;《BioMedical Engineering OnLine》;20170105;第1-17页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106960429A (en) | 2017-07-18 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106960429B (en) | A kind of CT image metal artifact bearing calibration and device | |
| Tian et al. | Low-dose CT reconstruction via edge-preserving total variation regularization | |
| EP3234919B1 (en) | System and method for image reconstruction | |
| La Riviere et al. | Reduction of noise-induced streak artifacts in X-ray computed tomography through spline-based penalized-likelihood sinogram smoothing | |
| JP6858664B2 (en) | Medical image processing equipment and medical diagnostic imaging equipment | |
| JP7330703B2 (en) | Medical image processing device and X-ray CT system | |
| JP6492005B2 (en) | X-ray CT apparatus, reconstruction calculation apparatus, and reconstruction calculation method | |
| Abdoli et al. | Metal artifact reduction strategies for improved attenuation correction in hybrid PET/CT imaging | |
| CN111047524A (en) | Low-dose CT lung image denoising method based on deep convolutional neural network | |
| CN105225208B (en) | A kind of computer tomography metal artifacts reduction method and device | |
| US20160078647A1 (en) | Metal artifacts reduction in cone beam reconstruction | |
| CN112381741B (en) | Tomography image reconstruction method based on SPECT data sampling and noise characteristics | |
| US20060285737A1 (en) | Image-based artifact reduction in PET/CT imaging | |
| US8855394B2 (en) | Methods and apparatus for texture based filter fusion for CBCT system and cone-beam image reconstruction | |
| WO2010096701A1 (en) | Projection-space denoising with bilateral filtering in computed tomography | |
| CN101541240A (en) | Computer tomography (CT) C-arm system and method for examination of an object | |
| CN105469366B (en) | A kind of analytic method of abatement CT image metal artifacts | |
| CN110599530B (en) | MVCT image texture enhancement method based on double regular constraints | |
| CN110458913B (en) | Method for correcting bone hardening artifacts in image reconstruction by multi-threshold segmentation CT image | |
| KR101824239B1 (en) | method and apparatus for reducing metal artifact | |
| EP3404618B1 (en) | Poly-energetic reconstruction method for metal artifacts reduction | |
| Wang et al. | Virtual colonoscopy screening with ultra low-dose CT and less-stressful bowel preparation: a computer simulation study | |
| CN111899312B (en) | Iterative compensation finite angle CT projection reconstruction method | |
| US20050018889A1 (en) | Systems and methods for filtering images | |
| Tang et al. | Efficient metal artifact reduction method based on improved total variation regularization |
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 |