CN103971349B - Computed tomography images method for reconstructing and ct apparatus - Google Patents
Computed tomography images method for reconstructing and ct apparatus Download PDFInfo
- Publication number
- CN103971349B CN103971349B CN201310037669.3A CN201310037669A CN103971349B CN 103971349 B CN103971349 B CN 103971349B CN 201310037669 A CN201310037669 A CN 201310037669A CN 103971349 B CN103971349 B CN 103971349B
- Authority
- CN
- China
- Prior art keywords
- image
- initial data
- computing unit
- rebuild
- isotropism
- 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
Landscapes
- Apparatus For Radiation Diagnosis (AREA)
Abstract
The invention discloses a kind of computed tomography images method for reconstructing and ct apparatus.The computed tomography images method for reconstructing includes：One first image I is rebuild according to initial data r0；According to initial data r and described first image I0One second image I is rebuild with alternative manner1；Calculate described first image I0With the second image I1Between poor Δ I=I0‑I1；Make described first image I0With the second image I1Between poor Δ I Δ I'=Δ I*h are obtained by an isotropism low pass filter, wherein h is the shock response of the isotropism low pass filter, and symbol * represents convolution；By I1(1 Δ I') projection is obtained to sinogram domainWherein symbol represents dot product；By I1Δ I' is projected to sinogram domain and is obtained r'；CalculateWith the difference between initial data rΔ r inverse projections to image area are obtained into Δ I "；From the second image I1Subtract Δ I " and obtain I2=I1‑ΔI″。
Description
Technical field
The present invention relates to Medical Imaging, particularly computed tomography.
Background technology
In computed tomography (CT), the X-ray after decaying through object is converted to electric signal and is generated by detector
Initial data, initial data is via a series of computing, and --- rebuilding --- obtains image.Image is converted to by initial data
Computing be called inverse projection, it is on the contrary then be projection.Because the frequency spectrum of initial data is similar to sine curve, therefore, image is become
Corresponding initial data is changed to be also referred to as projecting image onto sinogram domain.Compared to analytic reconstruction method, iterative reconstruction has
Many advantages, for example, can reduce pencil-beam artifact and metal artifacts.Pencil-beam artifact is mainly by the inaccurate of algorithm for reconstructing
Caused by property, metal artifacts are due to that X-ray is almost fully absorbed by metal, do not collect effectively believe on the detector
Caused by number.But, the speed of iterative approximation is slower.This shortcoming limits the application of iterative approximation.
Fig. 1 is the image after 1 iteration according to the iterative manner of prior art.This time iteration can be obtained based on analytic method
The image obtained.As shown in figure 1, there is artifact in the signified region of mark 102,104 and 106.
Fig. 2 is the image after successive ignition according to the iterative manner of prior art.What Fig. 2 and Fig. 1 was presented is identical to break
Layer, the image of same area.As shown in Fig. 2 the region of the meaning of mark 102,104 and 106 in original Fig. 1, is not deposited
In artifact.
At present, many methods are used for the speed for accelerating iterative approximation, accelerate convergence rate for example with regularization, make
Accelerate calculating speed with General Porcess Unit (GPU) and the calculating of projection and inverse projection is reduced using new mathematical method
Complexity.
The content of the invention
In view of this, the present invention proposes a kind of computed tomography images method for reconstructing and computed tomography is set
It is standby, to accelerate image reconstruction speed.
According to the first aspect of the invention there is provided a kind of computed tomography images method for reconstructing, including：Step
S302, one first image I is rebuild according to initial data r0；Step S304, according to initial data r and described first image I0With repeatedly
One second image I is rebuild for method1；Step S306, calculates described first image I0With the second image I1Between poor Δ I=I0-
I1；Step S308, makes described first image I0With the second image I1Between poor Δ I by an isotropism low pass filter
Δ I'=Δ I*h are obtained, wherein h is the shock response of the isotropism low pass filter, and symbol * represents convolution；Step
S310, by I1(1- Δ I') projection is obtained to sinogram domainWherein symbol represents dot product；Step S312, by I1·Δ
I' is projected to sinogram domain and is obtained r'；Step S314, is calculatedWith the difference between initial data rStep
S316, Δ I " is obtained by Δ r inverse projections to image area；Step S318, from the second image I1Subtract Δ I " and obtain I2=
I1-ΔI”。
In one embodiment of this invention, methods described can further comprise：Step S320, judges whether Δ r is less than r's
One preset ratio or step S312 to step S318 whether the preset times of executed one, if so, then terminating, if it is not, then performing
Step S322；Step S322, by I1It is updated to I2, and return to step S312 and its subsequent step.
In one embodiment of this invention, the preset ratio can be 1% to 3%, the preset times can for 3 to
5。
In one embodiment of this invention, it is described that one first image I is rebuild according to initial data r0Including according to original number
Described first image I is rebuild with analytic method according to r0。
In one embodiment of this invention, the analytic method is weighted filtering back projection.
In one embodiment of this invention, it is described that one first image I is rebuild according to initial data r0Including according to original number
Described first image I is rebuild with alternative manner according to r0。
According to the second aspect of the invention there is provided a kind of ct apparatus, including：One first computing unit,
It rebuilds one first image I according to initial data r0；One second computing unit, it is according to initial data r and described first image I0
One second image I is rebuild with alternative manner1；One the 3rd computing unit, it calculates described first image I0With the second image I1Between
Poor Δ I=I0-I1；One filter unit, it makes described first image I0With the second image I1Between poor Δ I by one it is each to
Same sex low pass filter and obtain Δ I'=Δ I*h, wherein h be the isotropism low pass filter shock response, symbol *
Represent convolution；One first projecting cell, it is by I1(1- Δ I') projection is obtained to sinogram domainWherein symbol is represented
Dot product；One the 4th computing unit, it is by I1Δ I' is projected to sinogram domain and is obtained r', is calculatedBetween initial data r
DifferenceΔ r inverse projections to image area are obtained into Δ I ", and from the second image I1Subtract Δ I " and obtain
I2=I1-ΔI”。
In one embodiment of this invention, the ct apparatus can further comprise：One judging unit, its
Judge Δ r whether be less than r a preset ratio or the 4th computing unit whether the preset times of executed one, if so, then
Terminate, if it is not, then calling an image update unit；Described image updating block is by I1It is updated to I2, and call the 4th calculating
Unit.
In one embodiment of this invention, the preset ratio can be 1% to 3%, the preset times can for 3 to
5。
In one embodiment of this invention, first computing unit is according to initial data r is rebuild with analytic method
First image I0。
In one embodiment of this invention, the analytic method is weighted filtering back projection.
In one embodiment of this invention, first computing unit is according to initial data r is rebuild with alternative manner
First image I0。
The computed tomography images method for reconstructing and ct apparatus of the present invention is by recognizing pseudo- shadow zone
Domain simultaneously only is iterated reconstruction to reduce operand to artifact region, and therefore, the speed of image reconstruction can substantially be accelerated.
Brief description of the drawings
The preferred embodiments of the present invention will be described in detail by referring to accompanying drawing below, and make one of ordinary skill in the art more
In the above and other feature and advantage of the clear present invention, accompanying drawing：
Fig. 1 is the image after 1 iteration according to the iterative manner of prior art.
Fig. 2 is the image after successive ignition according to the iterative manner of prior art.
Fig. 3 is that the flow of the computed tomography images method for reconstructing of the first embodiment according to the present invention is illustrated
Figure.
Fig. 4 is the block diagram of the ct apparatus of the second embodiment according to the present invention.
In above-mentioned accompanying drawing, the reference used is as follows：
102nd, 104,106 mark
300 computed tomography images method for reconstructing
400 ct apparatus
402 first computing units
404 second computing units
406 the 3rd computing units
408 filter units
410 first projecting cells
411 the 4th computing units
420 judging units
422 image update units
S302, S304, S306, S308, S310, S312 step
S314, S316, S318, S320, S322 step
Embodiment
To make the object, technical solutions and advantages of the present invention clearer, by the following examples to of the invention further detailed
Describe in detail bright.
The computed tomography images method for reconstructing and ct apparatus of the present invention is based on following observation：With
Sinogram domain is iterated reconstruction and is mainly used in reducing artifact, also, artifact only accounts for the sub-fraction of CT images.Therefore once exist
Image area detects artifact, then each iteration only needs to update artifact region.Can thus reduce projection and
The time of inverse projection.
Fig. 3 is that the flow of the computed tomography images method for reconstructing 300 of the first embodiment according to the present invention is shown
It is intended to.As shown in figure 3, computed tomography images method for reconstructing 300 includes：
Step S302：One first image I is rebuild according to initial data r0.In the present embodiment, can be according to initial data r
The first figure is rebuild with weighted filtering back projection (Weighted Filtered Back Projection) or other analytic methods
As I0.In other embodiments, can also be according to initial data r with the first image I of alternative manner reconstruction0。
Step S304：According to initial data r and the first image I0One second image I is rebuild with alternative manner1。
Step S306：Calculate the first image I0With the second image I1Between poor Δ I=I0-I1。
Step S308：Make the first image I0With the second image I1Between poor Δ I by an isotropism low pass filter
Δ I'=Δ I*h are obtained, wherein h is the shock response of the isotropism low pass filter, and symbol * represents convolution.H cutoff frequency
Rate can be determined according to actual needs.As other images, Δ I' is two-dimensional matrix.Values of the Δ I' at artifact-free position
For 0, there is the value at the position of artifact identical and non-zero (such as can be 1).Because in general artifact exists only in office
Portion region, therefore, the corresponding value of most point is 0 in Δ I', and in follow-up iteration, images of these points will be carried out no longer
Rebuild.This means that the quantity of the pixel used in projection and inverse projection or voxel will greatly reduce.Therefore, image reconstruction
Speed can substantially be accelerated.
Step S310：By I1(1- Δ I') projection is obtained to sinogram domainWherein symbol represents dot product.
Step S312：By I1Δ I' is projected to sinogram domain and is obtained r'.
Step S314：Calculate the difference between r'+r and initial data r
Step S316：Δ r inverse projections to image area are obtained into Δ I ".
Step S318：From the second image I1Subtract Δ I " and obtain I2=I1-ΔI”。
By from the second image I1Δ I " is subtracted, the artifact (such as pencil-beam artifact and metal artifacts) being originally present can subtract
It is few, therefore I2It is than the second image I1Better image.To obtain better image, above-mentioned step S312 can also be repeated extremely
Step S318.For example, computed tomography images method for reconstructing 300 may also include：
Step S320：Judge Δ r whether be less than r a preset ratio or step S312 to step S318 whether executed
One preset times, if so, then terminating, if it is not, then performing step S322；
Step S322：By I1It is updated to I2, and return to step S312 and its subsequent step.
Above-mentioned preset ratio can be 1% to 3%, such as 1%, 2%, 3%.Δ r is smaller, iteration production adjacent twice
Difference between raw image is also just smaller, therefore the iteration of more unnecessary progress again.Above-mentioned preset times can be 3
To 5, such as 3,4,5.Typically, iterative algorithm has a faster convergence rate, and iteration can ensure picture quality 3 to 5 times.
Fig. 4 is the block diagram of the ct apparatus 400 of the second embodiment according to the present invention.Saved in Fig. 4
The part unrelated with the improvement of the present invention is omited.As shown in figure 4, ct apparatus 400, which includes one first, calculates single
First 402, one second computing unit 404, one the 3rd computing unit 406, a filter unit 408, one first projecting cell 410 and one
4th computing unit 411.
First computing unit 402 rebuilds one first image I according to initial data r0.In the present embodiment, first calculate single
Member 402 can according to initial data r with weighted filtering back projection (Weighted Filtered Back Projection) or
Other analytic methods rebuild the first image I0.In other embodiments, the first computing unit 402 can also be according to initial data r
First image I is rebuild with alternative manner0。
Second computing unit 404 is according to initial data r and the first image I0One second image I is rebuild with alternative manner1。
3rd computing unit 406 calculates the first image I0With the second image I1Between poor Δ I=I0-I1。
Filter unit 408 makes the first image I0With the second image I1Between poor Δ I pass through an isotropism low pass filter
And Δ I'=Δ I*h are obtained, wherein h is the shock response of the isotropism low pass filter, and symbol * represents convolution.H cut-off
Frequency can be determined according to actual needs.As other images, Δ I' is two-dimensional matrix.Δ I' taking at artifact-free position
It is worth for 0, is having the value at the position of artifact identical and non-zero (such as can be 1).Because in general artifact is existed only in
Regional area, therefore, the corresponding value of most point is 0 in Δ I', and in follow-up iteration, images of these points will no longer enter
Row is rebuild.This means that the quantity of the pixel used in projection and inverse projection or voxel will greatly reduce.Therefore, image reconstruction
Speed can substantially accelerate.
First projecting cell 410 is by I1(1- Δ I') projection is obtained to sinogram domainWherein symbol is represented a little
Multiply.
4th computing unit 411 is by I1Δ I' is projected to sinogram domain and is obtained r', is calculatedWith initial data r it
Between differenceΔ r inverse projections to image area are obtained into Δ I ", and from the second image I1Subtract Δ I " and obtain I2
=I1-ΔI”。
By from the second image I1Δ I " is subtracted, the artifact (such as pencil-beam artifact and metal artifacts) being originally present can subtract
It is few, therefore I2It is than the second image I1Better image.To obtain better image, the 4th computing unit can also be repeated
411.For example, ct apparatus 400 may also include a judging unit 420 and an image update unit 422.
Judging unit 420 judge Δ r whether be less than r a preset ratio or the 4th computing unit 411 whether executed
One preset times, if so, then terminating, if it is not, then calling figure is as updating block 422.
Image update unit 422 is by I1It is updated to I2, and call the 4th computing unit 411.
Above-mentioned preset ratio can be 1% to 3%, such as 1%, 2%, 3%.Δ r is smaller, iteration production adjacent twice
Difference between raw image is also just smaller, therefore the iteration of more unnecessary progress again.Above-mentioned preset times can be 3
To 5, such as 3,4,5.Typically, iterative algorithm has a faster convergence rate, and iteration can ensure picture quality 3 to 5 times.
The computed tomography images method for reconstructing and ct apparatus of the present invention is by recognizing pseudo- shadow zone
Domain simultaneously only is iterated reconstruction to reduce operand to artifact region, and therefore, the speed of image reconstruction can substantially be accelerated.
The foregoing is merely illustrative of the preferred embodiments of the present invention, is not intended to limit the invention, all essences in the present invention
God is with principle, and any modification, equivalent substitution and improvements made etc. should be included in the scope of the protection.
Claims (12)
1. a kind of computed tomography images method for reconstructing, including：
Step S302：One first image I is rebuild according to initial data r0；
Step S304：According to initial data r and described first image I0One second image I is rebuild with alternative manner1；
Step S306：Calculate described first image I0With the second image I1Between poor Δ I=I0-I1；
Step S308：Make described first image I0With the second image I1Between poor Δ I by an isotropism low pass filter
Δ I'=Δ I*h are obtained, wherein h is the shock response of the isotropism low pass filter, and symbol * represents convolution；
Step S310：By I1(1- Δ I') projection is obtained to sinogram domainWherein symbol represents dot product；
Step S312：By I1Δ I' is projected to sinogram domain and is obtained r'；
Step S314：CalculateWith the difference between initial data r
Step S316：Δ r inverse projections to image area are obtained into Δ I "；
Step S318：From the second image I1Subtract Δ I " and obtain I2=I1-ΔI″。
2. the method as described in claim 1, it is characterized in that, methods described further comprises：
Step S320：Judge Δ r whether be less than r a preset ratio or step S312 to step S318 whether executed one is pre-
If number of times, if so, then terminating, if it is not, then performing step S322；
Step S322：By I1It is updated to I2, and return to step S312 and its subsequent step.
3. method as claimed in claim 2, it is characterized in that, the preset ratio is 1% to 3%, and the preset times are 3 to 5.
4. the method as described in claim 1, it is characterized in that, it is described that one first image I is rebuild according to initial data r0Including basis
Initial data r rebuilds described first image I with analytic method0。
5. method as claimed in claim 4, it is characterized in that, the analytic method is weighted filtering back projection.
6. the method as described in claim 1, it is characterized in that, it is described that one first image I is rebuild according to initial data r0Including basis
Initial data r rebuilds described first image I with alternative manner0。
7. a kind of ct apparatus, including：
One first computing unit, it rebuilds one first image I according to initial data r0；
One second computing unit, it is according to initial data r and described first image I0One second image I is rebuild with alternative manner1；
One the 3rd computing unit, it calculates described first image I0With the second image I1Between poor Δ I=I0-I1；
One filter unit, it makes described first image I0With the second image I1Between poor Δ I pass through an isotropism LPF
Device and obtain Δ I'=Δ I*h, wherein h is the shock response of the isotropism low pass filter, and symbol * represents convolution；
One first projecting cell, it is by I1(1- Δ I') projection is obtained to sinogram domainWherein symbol represents dot product；
One the 4th computing unit, it is by I1Δ I' is projected to sinogram domain and is obtained r', is calculatedBetween initial data r
DifferenceΔ r inverse projections to image area are obtained into Δ I'', and from the second image I1Subtract Δ I'' and obtain
Obtain I2=I1-ΔI''。
8. ct apparatus as claimed in claim 7, it is characterized in that, the ct apparatus enters one
Step includes：
One judging unit, its judge Δ r whether be less than r a preset ratio or the 4th computing unit whether executed one
Preset times, if so, then terminating, if it is not, then calling an image update unit；
Described image updating block is by I1It is updated to I2, and call the 4th computing unit.
9. ct apparatus as claimed in claim 8, it is characterized in that, the preset ratio is 1% to 3%, described
Preset times are 3 to 5.
10. ct apparatus as claimed in claim 7, it is characterized in that, first computing unit is according to original
Data r rebuilds described first image I with analytic method0。
11. ct apparatus as claimed in claim 10, it is characterized in that, the analytic method is that weighted filtering is inverse
Sciagraphy.
12. ct apparatus as claimed in claim 7, it is characterized in that, first computing unit is according to original
Data r rebuilds described first image I with alternative manner0。
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310037669.3A CN103971349B (en) | 2013-01-30 | 2013-01-30 | Computed tomography images method for reconstructing and ct apparatus |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310037669.3A CN103971349B (en) | 2013-01-30 | 2013-01-30 | Computed tomography images method for reconstructing and ct apparatus |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103971349A CN103971349A (en) | 2014-08-06 |
| CN103971349B true CN103971349B (en) | 2017-08-08 |
Family
ID=51240798
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201310037669.3A Active CN103971349B (en) | 2013-01-30 | 2013-01-30 | Computed tomography images method for reconstructing and ct apparatus |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103971349B (en) |
Families Citing this family (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2017006764A1 (en) * | 2015-07-08 | 2017-01-12 | 株式会社日立製作所 | Image computing device, image computing method, and tomograph |
| WO2017096609A1 (en) | 2015-12-11 | 2017-06-15 | Shanghai United Imaging Healthcare Co., Ltd. | System and method for image reconstruction |
| CN106373165B (en) * | 2016-08-31 | 2017-11-28 | 广州华端科技有限公司 | Tomography composograph method for reconstructing and system |
| CN107886478B (en) * | 2017-09-22 | 2020-10-30 | 深圳先进技术研究院 | CT image reconstruction method and system, terminal and readable storage medium |
| CN111096761B (en) * | 2018-10-29 | 2024-03-08 | 上海西门子医疗器械有限公司 | Method, device and related equipment for correcting scattering of wedge-shaped filter |
| CN109523605A (en) * | 2018-11-29 | 2019-03-26 | 上海联影医疗科技有限公司 | A kind of method, apparatus, equipment and the medium of CT image reconstruction |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1416781A (en) * | 2002-12-30 | 2003-05-14 | 北京质子科技开发有限公司 | Fast iterative reconstruction method of high-resolution heat source image in positive-electron tomographic scanning |
-
2013
- 2013-01-30 CN CN201310037669.3A patent/CN103971349B/en active Active
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1416781A (en) * | 2002-12-30 | 2003-05-14 | 北京质子科技开发有限公司 | Fast iterative reconstruction method of high-resolution heat source image in positive-electron tomographic scanning |
Non-Patent Citations (2)
| Title |
|---|
| Sparse Angular X-ray Cone Beam CT Image Iterative Reconstruction Using Normal-dose Scan Induced Nonlocal Prior;Hua Zhang等;《2012 IEEE Nuclear Science Symposium and Medical Imaging Conference Record》;20121103;第3671-3674页 * |
| 同步辐射光成像中的最大似然迭代重建法;张书铭 等;《CT理论与应用研究》;20070630;第16卷(第2期);第8-13页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN103971349A (en) | 2014-08-06 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103971349B (en) | Computed tomography images method for reconstructing and ct apparatus | |
| JP5315157B2 (en) | Information processing apparatus, line noise reduction processing method, and program | |
| JP5610761B2 (en) | X-ray image processing apparatus, X-ray image processing system, X-ray image processing method, and computer program | |
| US20160070979A1 (en) | Method and Apparatus for Generating Sharp Image Based on Blurry Image | |
| Chen et al. | CT metal artifact reduction method based on improved image segmentation and sinogram in‐painting | |
| CN102722875B (en) | Visual-attention-based variable quality ultra-resolution image reconstruction method | |
| CN103136731B (en) | A kind of parameter imaging method of dynamic PET images | |
| US9613405B2 (en) | Scalable massive parallelization of overlapping patch aggregation | |
| Al‐Ameen et al. | A new algorithm for improving the low contrast of computed tomography images using tuned brightness controlled single‐scale Retinex | |
| CN104751429A (en) | Dictionary learning based low-dosage energy spectrum CT image processing method | |
| CN103440676A (en) | Method for reconstruction of super-resolution coronary sagittal plane image of lung 4D-CT image based on motion estimation | |
| CN110246090B (en) | Image processing method and device | |
| US7826678B2 (en) | Adaptive image sharpening method | |
| CN110070510A (en) | A kind of CNN medical image denoising method for extracting feature based on VGG-19 | |
| CN107424121B (en) | Fuzzy image super-resolution reconstruction method based on fractional order differential | |
| Liu et al. | Learning noise-decoupled affine models for extreme low-light image enhancement | |
| CN109447935B (en) | Infrared image processing method and device, computer equipment and readable storage medium | |
| CN109118439B (en) | Cone beam CT deblurring method based on line integral | |
| CN106033598B (en) | Digital grid imaging method for eliminating scattered radiation influence | |
| CN105741255B (en) | Image interfusion method and device | |
| CN103559690B (en) | Method for achieving smoothness of image edges | |
| CN111080736B (en) | Low-dose CT image reconstruction method based on sparse transformation | |
| Bertaccini et al. | An adaptive norm algorithm for image restoration | |
| CN110443760A (en) | Image denoising method and computer readable storage medium | |
| Huo et al. | Removing ring artefacts in CT images via unidirectional relative variation model |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |