CN103714560B - A kind of image rebuilding method based on Katsevich algorithm and system - Google Patents
A kind of image rebuilding method based on Katsevich algorithm and system Download PDFInfo
- Publication number
- CN103714560B CN103714560B CN201310734168.0A CN201310734168A CN103714560B CN 103714560 B CN103714560 B CN 103714560B CN 201310734168 A CN201310734168 A CN 201310734168A CN 103714560 B CN103714560 B CN 103714560B
- Authority
- CN
- China
- Prior art keywords
- data
- projection
- image
- interpolation
- described step
- 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.)
- Expired - Fee Related
Links
Landscapes
- Apparatus For Radiation Diagnosis (AREA)
- Image Processing (AREA)
Abstract
The present invention proposes a kind of image rebuilding method based on Katsevich algorithm, by the error of each operation module in Katsevich algorithm for reconstructing is optimized, the impact on rebuilding image of the error artifact can be effectively reduced, and the information of original true CT image can be kept.Prove through repeatedly simulation practice, the method of the present invention is simple to operate, can effectively reduce the error artifact in CT image, the reconstruction image obtained by the method for the present invention can not only reduce the metal artifacts impact on image, and can keep the detailed information of area-of-interest.
Description
Technical field
The invention belongs to the image processing field of medical image, be specifically related to a kind of based on Katsevich algorithm pattern picture weight
The method and system built.
Background technology
In the reconstruction image of exact reconstruction algorithm katsevich, it may appear that various difform artifacts, these artifacts
Have impact on the accuracy of algorithm imaging.Wherein owing to the data for projection during rebuilding is carried out artifact produced by discretization just
It it is the most important a kind of source of artifact.Directly result in that to change artifact Producing reason mainly Katsevich algorithm itself be a kind of
Exact reconstruction algorithm, but in actual application or the data for projection that lacks is discrete, follow-up operation is all to dispersion number
According to having carried out corresponding approximation operation, thus result in final imaging containing substantial amounts of artifact.High-quality in order to obtain
CT rebuilds image, needs the error artifact that the discrete operations as far as possible reducing data for projection is brought, the method pretreatment generally used
The method optimized, cardinal principle is to using different derivative algorithms that data for projection is carried out derivation module in image reconstruction process
It is optimized, thus reaches to reduce the effect of artifact.
Rebuilding, from exact reconstruction algorithm Katsevich, the mechanism that image artifacts produces, multiple CT artifact pretreatment disappears
Except method is suggested in succession.The method of pretreatment mainly includes using a kind of higher Method of Seeking Derivative of precision to carry out derivation etc..Use
Single method carries out derivation, and the result being easily caused derivation can be lost, thus can produce new error, so weight to image
Build effect improved the most obvious.
Summary of the invention
In order to solve the deficiencies in the prior art, the present invention provide a kind of simple to operate, reconstructed image quality is high and rebuilds
Fireballing image rebuilding method based on Katsevich algorithm and system.
The present invention is achieved through the following technical solutions:
A kind of image rebuilding method based on Katsevich algorithm, comprises the following steps:
(1) data for projection rebuild required for image is obtained；
(2) described data for projection is carried out mixing derivation based on line-of-sight course and Richardson extrapolation, particularly as follows: to projection
Each row, column of data, frame start most and the derivation of last at least two data for projection use line-of-sight course derivation；Right
Data for projection in other uses Richardson extrapolation to carry out derivation, additionally carries out this module at fixed point in concrete operations
Reason；
(3) data for projection processing described step (2) carries out length weight correction；
(4) interpolation operation before the data for projection processing described step (3) is filtered, sits data for projection
The conversion of mark system, in this step, the married operation closest to interpolation Yu cubic convolution interpolation that data for projection is used, additionally exist
Concrete operations carry out fixed point process to this module;
(5) data for projection processing described step (4) carries out Hilbert transform, then carries out adding respective window function
Operation；
(6) data for projection processing described step (5) carries out interpolation, and data for projection carries out coordinate system change again
Change, recover the coordinate system of data for projection, additionally in concrete operations, this module is carried out fixed point process;
(7) data for projection processing described step (6) carries out back projection's generation reconstruction image, in this step to throwing
When shadow data carry out cumulative, the method using mixing difference to choose data for projection, additionally in concrete operations, this module is carried out
Fixed point processes.
Further, when the data for projection obtained contains noise, described step (1) also includes obtained projection number
According to being filtered process, to reduce noise in subsequent operation to rebuilding the impact of artifact in image.
Further, when described step (2) carries out derivation to data for projection, the method for employing is the method for mixing derivation,
That is: data acquisition difference coefficient, line-of-sight course or the five-spot derivation to edge, for middle data acquisition center difference coefficient, Li Chasen
Extrapolation or fitting process derivation.
Further, described step (5) uses fast Fourier transform FFT to complete Hilbert transform.Have only to into
FFT computing of row, and operation result is temporarily stored for subsequent calculations.
Further, described step (6) uses interpolation method to be the hybrid interpolation closest to interpolation Yu cubic convolution interpolation
Method.
Further, the method that the mixing difference in described step (7) chooses data for projection, particularly as follows: when coordinate (u*,
When w*) being positioned at the outermost of detecting plate, closest difference is used to choose the data for projection participating in integration；When coordinate (u*, w*) position
Except time on the intersection point of the rib in upper step and rib on detecting plate, use bilinear interpolation method and choose data for projection；When
Coordinate (u*, w*) is positioned on detecting plate except, time on the rib in above-mentioned two steps, the method using closest interpolation chooses throwing
Shadow data；When coordinate (u*, w*) is positioned on detecting plate the region except above-mentioned three steps, the method for closest interpolation is used to choose
Data for projection.
Further, when described step (7) obtains integrating range [sb, st] required during back projection, need completing
Solving of nonlinear equation, uses interative computation based on two way classification or Newton iteration method or Secant Method when solving.
Further, the method carrying out fixed point process in described step (2), (4), (6), (7), particularly as follows: use
Cordic algorithm surmounts function linearisation in solving to implement, and realizes using in pipeline organization 14 for cordic algorithm
Level flowing water progression, it is possible to meet the required precision of medical image, it addition, in fixed point, surface analysis devises based on affine
The algorithm of form realizes；Precision analysis devises algorithm based on quantization error propagation model and simulation self adaptation annealing algorithm phase
It is implemented in combination with;Meanwhile, according to different constraints, respectively obtaining range of error is 2-4, tri-groups of precision analysis of 2-8,2-13
Result.
Another aspect of the present invention, it is provided that a kind of image reconstruction side based on Katsevich algorithm performing the present invention
The system of method, it is characterised in that: this system includes that CT machine, image reconstruction work station, high-definition image display device, the X of CT machine penetrate
Detecting plate in ray tube, CT machine;Its working method is: the relative invariant position of the X-ray tube in CT machine and detecting plate, they
Rotating along scanning gantry 360 ° and be irradiated reconstruction object, the data for projection obtained is by PCI-E bus on a frame-by-frame basis
It is transferred in image reconstruction work station, image reconstruction work station utilizes the image based on Katsevich algorithm of the present invention
Method for reconstructing carries out the most quickly rebuilding, and the reconstruction image obtained is shown by high-definition image display device.
The invention has the beneficial effects as follows: the present invention is by carrying out the error of modules in Katsevich algorithm for reconstructing
Optimize, it is possible to be effectively reduced the impact on rebuilding image of the error artifact, correlation module has been carried out fixed point process simultaneously and has added
The fast speed rebuild, and the information of original true CT image can be kept.Prove through repeatedly simulation practice, the side of the present invention
Method is simple to operate, it is possible to effectively reduce the error artifact in CT image, by the reconstruction image of the method acquisition of the present invention not only
The metal artifacts impact on image can be reduced, and the detailed information of area-of-interest can be kept.
Accompanying drawing explanation
Fig. 1 is the image rebuilding method flow chart based on Katsevich algorithm of the present invention；
Fig. 2 is the method flow diagram that fixed point processes；
Fig. 3 is the detailed annotation figure of mixing derivation operations；
Fig. 4 is the interpolation schematic diagram of individual frames data for projection；
The middle hybrid interpolation schematic diagram of Tu5Shi back projection operation；
Fig. 6 is carried out the system cross sectional representation in the Y direction of the inventive method；
Fig. 7 (a) is the Shepp-Logan body mould of standard, is used as the ideal body mould of CT image reconstruction；
Fig. 7 (b) is the image not carrying out the Katsevich method for reconstructing reconstruction that invention optimizes；
Fig. 7 (c) is by the image rebuild based on the rear Katsevich method for reconstructing that the present invention optimizes；
In figure, the implication of each sign is: 1-CT machine, 2-image reconstruction work station, 3-high-definition image display device, 4-CT machine
X-ray bulb, detecting plate in 5-CT machine.
Detailed description of the invention
The present invention is further described for explanation and detailed description of the invention below in conjunction with the accompanying drawings.
Image rebuilding method based on Katsevich algorithm as shown in Figure 1, in turn includes the following steps:
(1) obtain CT machine systematic parameter and rebuild the data for projection required for image.When the data for projection obtained contains
During noise, need to be filtered obtained data for projection processing, to reduce noise in subsequent operation to rebuilding in image
The impact of artifact.
(2) data for projection obtaining step (1) carries out mixing derivation based on line-of-sight course and Richardson extrapolation, specifically
Embodiment as shown in Figure 3, the later last at least two started most of each row, column of data for projection, frame is thrown
The derivation of shadow data (MARG) uses line-of-sight course derivation；Data for projection for other uses Richardson extrapolation to ask
Lead.The main purpose using line-of-sight course is that it can with the derivative obtaining MARG of degree of precision, thus be avoided throwing
The most obtainable situation of shadow data derivative；And the purpose using Richardson extrapolation to carry out derivation be utilize Richardson extrapolation with
The advantage of the change derivation precision change extrapolation number of times requirement as the case may be can use different extrapolation situations
Reason looks into Senn process, meets the required precision of derivation to greatest extent.Be can be obtained by and original projection number by operation above
According to the data for projection after the derivation that amount is equal, on the one hand improve the derivation precision of data to greatest extent, on the other hand eliminate
After derivation, the reconstruction efficiency lost to improve algorithm for reconstructing of data, has carried out pinpointing the optimisation strategy of streamline to this module
This module is optimized.
(3) data for projection processing step (2) carries out length weight correction；
(4) interpolation operation before the data for projection processing step (3) is filtered, carries out coordinate system to data for projection
Conversion, in this step, the married operation closest to interpolation Yu cubic convolution interpolation using data for projection, such as accompanying drawing 4 institute
Show, be i.e. closest to interpolation to edge data acquisition, and cubic convolution interpolation is used for non-edge data.Due to
Precision closest to interpolation is the lowest so that rebuild image effect very poor, but simple to operate, bilinear interpolation can produce low pass
The effect of wave filter and to make to rebuild image fuzzyyer, and cubic convolution interpolation can both improve precision and also be able to avoid simultaneously
Obscuring of image, therefore improves the computational accuracy of interpolating module, and cubic convolution interpolation makes computing quantitative change big simultaneously.In order to carry
The reconstruction efficiency of high algorithm for reconstructing, this module is optimized by the optimisation strategy that this module has carried out pinpointing streamline, can
So that this module area of whole hardware circuit when carrying out hardware and realizing is less, faster, power consumption is lower for speed.
(5) data processing step (4) carry out Hilbert transform, and usual data for projection is to be come by convolution operation
Completing Hilbert transform, in the present invention, the fast Fourier transform (FFT) of employing completes Hilbert transform, complete
After having become Hilbert transform, the data for projection processed is added respective window function operation.
(6) data for projection processing step (5) carries out interpolation, and data for projection carries out coordinate system transformation again,
Recover the coordinate system of data for projection, remain the married operation closest to interpolation Yu cubic convolution interpolation what this step used.
Module is optimized by the streamline simultaneously using fixed point.
(7) data for projection producing step (6) carries out back projection's generation reconstruction image, in this step to data for projection
When carrying out cumulative, the present invention uses and a kind of mixes the method that difference chooses data for projection, as shown in Figure 5: work as coordinate
When (u*, w*) is positioned at the outermost of detecting plate, the present invention uses closest difference to choose the data for projection participating in integration；Work as coordinate
When (u*, w*) is positioned on detecting plate on the intersection point except the rib in upper step and rib, uses bilinear interpolation method and choose throwing
Shadow data；When coordinate (u*, w*) is positioned on detecting plate except time on the rib in above-mentioned two steps, use the side of closest interpolation
Method chooses data for projection；When coordinate (u*, w*) is positioned on detecting plate the region except above-mentioned three steps, use closest interpolation
Method chooses data for projection, so that precision is higher when choosing data for projection and being updated reconstruction point.In addition to add
The fast speed rebuild, the present invention has carried out fixed point streamline optimization to this module, so that rebuild in hgher efficiency, power consumption is more
Low.
When step (2) carries out derivation to the data for projection acquired in step (1), the method using mixing derivation: to edge
Data can use difference coefficient, line-of-sight course or five-spot etc., can use, for middle data, the Method of Seeking Derivative that precision is higher,
The such as method of center difference coefficient, Richardson extrapolation or fitting derivative so that the data volume after derivation and derivation it
Before holding consistent, it is to avoid the loss of data and error therefore.
When step (5) carries out FFT computing to the data in step (4), owing to needing repeatedly to use the FFT of Hilbert core
Operation result, in order to accelerate computing in process of reconstruction, it is only necessary to it is carried out once-through operation, and is deposited by its operation result temporarily
Storage is for subsequent calculations.
Fixed point process method as shown in Figure 2, particularly as follows: use cordic algorithm solve to implement in surmount
Function linearization, realizes using in pipeline organization 14 grades of flowing water progression for cordic algorithm, it is possible to meet medical image
Required precision, it addition, in fixed point, surface analysis devises algorithm based on affine form and realizes；Precision analysis devises
Algorithm based on quantization error propagation model and simulation self adaptation annealing algorithm combine realization;Meanwhile, according to different constraints
Condition, respectively obtaining range of error is 2-4, the result of tri-groups of precision analysis of 2-8,2-13
As shown in Figure 6, being carried out the system of the image rebuilding method based on Katsevich algorithm of the present invention, this is
System includes the detecting plate in CT machine, image reconstruction work station, high-definition image display device, the X-ray bulb of CT machine, CT machine;Its
Working method is: the relative invariant position of the X-ray tube in CT machine and detecting plate, and they rotate counterweight along scanning gantry 360 °
Building object to be irradiated, the data for projection obtained on a frame-by-frame basis is transferred in image reconstruction work station by PCI-E bus,
The image rebuilding method based on Katsevich algorithm utilizing the present invention in image reconstruction work station carries out the most quickly rebuilding,
The reconstruction image obtained is shown by high-definition image display device.
As shown in Figure 7, Fig. 7-a is the Shepp-Logan body mould of standard, is used as the ideal body mould of CT image reconstruction,
Fig. 7-b is the image not carrying out the Katsevich method for reconstructing reconstruction that invention optimizes, and Fig. 7-c is by based on optimization of the present invention
Rear Katsevich method for reconstructing rebuild image, by the mean square deviation of image and rectangular histogram are carried out right, after optimization figure
Significantly declining as mean square deviation has had, the most histogrammic high-frequency noise is also significantly improved.
Above content is to combine concrete preferred implementation further description made for the present invention, it is impossible to assert
Being embodied as of the present invention is confined to these explanations.For general technical staff of the technical field of the invention,
On the premise of present inventive concept, it is also possible to make some simple deduction or replace, all should be considered as belonging to the present invention's
Protection domain.
Claims (10)
1. an image rebuilding method based on Katsevich algorithm, it is characterised in that said method comprising the steps of:
(1) data for projection rebuild required for image is obtained；
(2) described data for projection is carried out mixing derivation based on line-of-sight course and Richardson extrapolation, particularly as follows: to data for projection
Each row, column, frame start most and the derivation of last at least two data for projection use line-of-sight course derivation；For it
His data for projection uses Richardson extrapolation to carry out derivation, uses fixed point to process this step in concrete operations；
(3) data for projection processing described step (2) carries out length weight correction；
(4) interpolation operation before the data for projection processing described step (3) is filtered, carries out coordinate system to data for projection
Conversion, in this step, the married operation closest to interpolation Yu cubic convolution interpolation using data for projection, in concrete operations
In to this step use fixed point process；
(5) data for projection processing described step (4) carries out Hilbert transform, then carries out adding respective window function behaviour
Make；
(6) data for projection processing described step (5) carries out interpolation, and data for projection carries out coordinate system transformation again,
Recover the coordinate system of data for projection, use fixed point to process this step in concrete operations；
(7) data for projection processing described step (6) carries out back projection's generation reconstruction image, in this step to projection number
According to when carrying out cumulative, use and mix the method that difference chooses data for projection, in concrete operations, this step is used at fixed point
Reason.
Method the most according to claim 1, it is characterised in that: when the data for projection obtained contains noise, described step
(1) also include being filtered obtained data for projection processing, to reduce noise in subsequent operation to rebuilding puppet in image
The impact of shadow.
Method the most according to claim 1, it is characterised in that: when described step (2) carries out derivation to data for projection, use
Method be the method for mixing derivation, it may be assumed that data acquisition difference coefficient, line-of-sight course or the five-spot derivation to edge, for middle
Data acquisition center difference coefficient, Richardson extrapolation or fitting process derivation.
Method the most according to claim 1, it is characterised in that: described step (5) uses fast Fourier transform FFT to come
Become Hilbert transform.
Method the most according to claim 4, it is characterised in that: have only to carry out a FFT computing, and operation result is faced
Time storage for subsequent calculations.
Method the most according to claim 1, it is characterised in that: described step (6) uses interpolation method to be closest to interpolation
Hybrid interpolation method with cubic convolution interpolation.
Method the most according to claim 1, it is characterised in that: the mixing difference in described step (7) chooses data for projection
Method, particularly as follows: when coordinate (u*, w*) is positioned at the outermost of detecting plate, uses closest difference to choose and participate in integration
Data for projection；When coordinate (u*, w*) is positioned on detecting plate on the intersection point except the rib in upper step and rib, use bilinearity
Interpolation method chooses data for projection；When coordinate (u*, w*) is positioned on detecting plate except time on the rib in above-mentioned two steps, use
The method of closest interpolation chooses data for projection；When coordinate (u*, w*) is positioned on detecting plate the region except above-mentioned three steps, adopt
Data for projection is chosen by the method for closest interpolation.
Method the most according to claim 1, it is characterised in that: described step obtains integration required during back projection in (7)
Time interval [sb, st], need to solve completing nonlinear equation, when solving, use interative computation based on two way classification or newton
Iterative method or Secant Method.
Method the most according to claim 1, it is characterised in that: described step (2), (4), (6), (7) carry out fixed point
The method processed, particularly as follows: use cordic algorithm to surmount function linearisation, for cordic algorithm in solving to implement
Realize pipeline organization uses 14 grades of flowing water progression, it is possible to meet the required precision of medical image, it addition, in fixed point,
Surface analysis devises algorithm based on affine form and realizes；Precision analysis devises algorithm based on quantization error propagation model
Combine realization with simulation self adaptation annealing algorithm；Meanwhile, according to different constraints, respectively obtaining range of error is 2-4,
The result of tri-groups of precision analysis of 2-8,2-13.
10. the image rebuilding method based on Katsevich algorithm performed as described in any one of claim 1-8 is
System, it is characterised in that: this system include CT machine, image reconstruction work station, high-definition image display device, the X-ray bulb of CT machine,
Detecting plate in CT machine；Its working method is: the relative invariant position of the X-ray bulb in CT machine and detecting plate, they along
Scanning gantry 360 ° rotates and is irradiated reconstruction object, and the data for projection obtained on a frame-by-frame basis is transmitted by PCI-E bus
In image reconstruction work station, utilize in image reconstruction work station as described in any one of claim 1-8 based on
The image rebuilding method of Katsevich algorithm carries out accurately quickly rebuilding, and the reconstruction image obtained is shown by high-definition image and sets
For showing.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310734168.0A CN103714560B (en) | 2013-12-27 | 2013-12-27 | A kind of image rebuilding method based on Katsevich algorithm and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310734168.0A CN103714560B (en) | 2013-12-27 | 2013-12-27 | A kind of image rebuilding method based on Katsevich algorithm and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103714560A CN103714560A (en) | 2014-04-09 |
| CN103714560B true CN103714560B (en) | 2016-10-05 |
Family
ID=50407501
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201310734168.0A Expired - Fee Related CN103714560B (en) | 2013-12-27 | 2013-12-27 | A kind of image rebuilding method based on Katsevich algorithm and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103714560B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105809723B (en) * | 2016-03-09 | 2018-10-19 | 哈尔滨工业大学深圳研究生院 | CBCT method for reconstructing and system |
| CN106056646B (en) * | 2016-05-24 | 2018-02-23 | 广州华端科技有限公司 | CT image rebuilding methods and system |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1788684A (en) * | 2004-09-29 | 2006-06-21 | 株式会社东芝 | Image reconstruction method and x-ray computed tomography apparatus |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7539281B2 (en) * | 2004-10-15 | 2009-05-26 | General Electric Company | Methods and apparatus for reconstruction in helical cone beam volumetric CT |
-
2013
- 2013-12-27 CN CN201310734168.0A patent/CN103714560B/en not_active Expired - Fee Related
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1788684A (en) * | 2004-09-29 | 2006-06-21 | 株式会社东芝 | Image reconstruction method and x-ray computed tomography apparatus |
Non-Patent Citations (4)
| Title |
|---|
| Studies on implementation of the Katsevichalgorithm for spiral cone-beam CT;Hengyong Yua et.al.;《Journal of X-Ray Science and Technology》;20041231;第12卷(第2期);97-116 * |
| 圆弧轨迹Katsevich类型的锥束CT重建;王瑜等;《光电子激光》;20080731;第19卷(第7期);967-971 * |
| 螺旋锥束ＣＴ三维图像重建算法中求导模块的设计与实现;贾喜元;《中国优秀硕士学位论文全文数据库_信息科技辑》;20111215;I138-1510 * |
| 螺旋锥束工业CT的Katsevich精确重建算法研究;牛小明;《中国优秀硕士学位论文全文数据库_信息科技辑》;20090615;I138-768 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN103714560A (en) | 2014-04-09 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Kurz et al. | CBCT correction using a cycle-consistent generative adversarial network and unpaired training to enable photon and proton dose calculation | |
| Yu et al. | Deep sinogram completion with image prior for metal artifact reduction in CT images | |
| Landry et al. | Comparing Unet training with three different datasets to correct CBCT images for prostate radiotherapy dose calculations | |
| CN103279929B (en) | A kind of prediction of the CT image metallic traces based on integral cosine and artifact minimizing technology | |
| CN102314698B (en) | Total variation minimization dosage CT (computed tomography) reconstruction method based on Alpha divergence constraint | |
| CN102945328B (en) | Based on the x-ray imaging image simulation method of GPU concurrent operation | |
| CN112102213B (en) | Low-dose CT image processing method, scanning system and computer storage medium | |
| CN104268914B (en) | Reestablishing method of 4D-CT (Four Dimensional-Computed Tomography) different time phase sequence image | |
| CN103295234B (en) | Based on the medical image segmentation system and method for deformation surface model | |
| CN103810734B (en) | A kind of low dose X-ray CT data for projection restoration methods | |
| CN103247061A (en) | Augmented lagrangian iterative reconstruction method of X-ray image and CI image | |
| CN106127825B (en) | A kind of X ray CT image rebuilding method based on broad sense punishment weighted least-squares | |
| WO2024051018A1 (en) | Pet parameter image enhancement method and apparatus, device, and storage medium | |
| CN105931280A (en) | GPU-based fast three-dimensional CT (computed tomography) iterative reconstruction method | |
| CN103793890A (en) | Method for recovering and processing energy spectrum CT images | |
| CN103714560B (en) | A kind of image rebuilding method based on Katsevich algorithm and system | |
| Dillon et al. | Evaluating reconstruction algorithms for respiratory motion guided acquisition | |
| Zhang et al. | 3D remote healthcare for noisy CT images in the Internet of things using edge computing | |
| CN103226815A (en) | Low dose CT image filtering method | |
| CN102024267A (en) | Low-dose computed tomography (CT) image processing method based on wavelet space directional filtering | |
| CN103340602B (en) | Method for optimizing optimal viewing angle of multiple branch interesting blood vessel section | |
| Li et al. | MARGANVAC: metal artifact reduction method based on generative adversarial network with variable constraints | |
| CN102243766A (en) | New method for accelerating reconstruction of computerized tomography (CT) image | |
| Yang et al. | Cycle-consistent learning-based hybrid iterative reconstruction for whole-body PET imaging | |
| Sun et al. | ACnerf: enhancement of neural radiance field by alignment and correction of pose to reconstruct new views from a single x-ray |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20161005 Termination date: 20201227 |