CN100444798C - Rotary symmetric voxel discretization method for CT reestablishing - Google Patents

Rotary symmetric voxel discretization method for CT reestablishing Download PDF

Info

Publication number
CN100444798C
CN100444798C CNB2005100122508A CN200510012250A CN100444798C CN 100444798 C CN100444798 C CN 100444798C CN B2005100122508 A CNB2005100122508 A CN B2005100122508A CN 200510012250 A CN200510012250 A CN 200510012250A CN 100444798 C CN100444798 C CN 100444798C
Authority
CN
China
Prior art keywords
projection
layer
annulus
discretization
calculating
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CNB2005100122508A
Other languages
Chinese (zh)
Other versions
CN1899219A (en
Inventor
邢宇翔
陈志强
张丽
李元景
刘以农
赵自然
肖永顺
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Tsinghua University
Nuctech Co Ltd
Original Assignee
Tsinghua University
Nuctech Co Ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Tsinghua University, Nuctech Co Ltd filed Critical Tsinghua University
Priority to CNB2005100122508A priority Critical patent/CN100444798C/en
Publication of CN1899219A publication Critical patent/CN1899219A/en
Application granted granted Critical
Publication of CN100444798C publication Critical patent/CN100444798C/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Abstract

The present invention relates to CT image reconstructing technology, and is especially one rotary symmetric voxel discretizing method for CT reconstruction. The method includes the following steps: 1. calculating the sampling step pitch d based on the reconstructing resolution; 2. discretizing the effective reconstructing height in the scan rotating axis direction into discs of thickness d; 3. discretizing each of the scan discs in the scan rotary radius direction into rings; 4. calculating the voxel numbers in each layer; 5. discretizing each ring in circumference direction; 6. calculating system matrix parameters corresponding to any projection angle; 7. calculating projection/back projection in table look up method; and 8. interpolating the reconstructed image based on the display mode. The method can reduce the required calculation amount and memory space of system matrix coefficients and raise CT reconstruction speed while ensuring the required resolution.

Description

A kind of plain discretization method of rotationally symmetric body that is used for the CT reconstruction
Technical field
The present invention relates to fault image reconstruction technique field, the plain discretization method of rotationally symmetric body in rebuilding especially for CT.
Background technology
Along with the progress in epoch, computer tomography technology (CT) has obtained using widely in medical diagnostic techniqu and Non-Destructive Testing.In the CT algorithm for reconstructing, amount of calculation mainly concentrates on projection/back projection's step.Under the certain situation of CT scan systematic parameter, calculate the sytem matrix of using in projection/back projection in advance, realize that with look-up table projection/back projection is the reconstruction accelerated method of using always.The statistics iterative method has remarkable advantages in the calculating of reconstructed image, but repeatedly the required amount of calculation of iteration is very huge, and the normal look-up table that uses can be avoided the double counting of each iterative process to sytem matrix in iterative method.Simultaneously, look-up table also is applicable to the fixed engineering practical application of systematic parameter.Look-up table has proposed the demand to internal memory when reducing amount of calculation, along with the growth of unknown space dimensionality, the required space of storage system matrix exceeds the capacity of present universal computing platform with the sharp increase of power power.So the use of look-up table at present only limits to wisp or low-resolution image is rebuild.In sum, how to construct matrix elements table efficiently, reducing required memory headroom is the key that look-up table is applied to high-resolution and high dimension image reconstruction, also is the important foundation that the statistics iterative method obtains broader applications in practice.The pole of rotation coordinate method is a kind of method of saving the look-up table required memory and improving the speed of tabling look-up that people propose, but stoped the generally use of this method along with the increase resolution of rebuilding radius can descend, and the equal difference ranking method of this being done behind the method improvement obviously increases the complexity of calculating.
Summary of the invention
In order to overcome the deficiency that above-mentioned existing method exists, the purpose of this invention is to provide the plain discretization method of rotationally symmetric body that a kind of CT of being used for rebuilds.It can reduce required sytem matrix coefficient calculations amount and the memory space of projection/back projection in the CT algorithm for reconstructing under the prerequisite that guarantees resolution, improve the CT reconstruction speed.
In order to reach above-mentioned goal of the invention, technical scheme of the present invention realizes as follows:
A kind of plain discretization method of rotationally symmetric body that is used for the CT reconstruction, its key step is:
1) requires calculating sampling step pitch d according to what rebuild resolution;
2) axis direction of edge scanning rotating shaft will effectively be rebuild and highly be separated into the disk that thickness is d;
3) every layer of disk scanned radius of turn direction discretization in diameter of section is the effective coverage of Nd, be separated into the annulus of N/2+1 ring width d;
4) the voxel number in calculating every layer
Figure C20051001225000051
Θ is the projection number in the 360 degree scopes;
5) in every layer cross section, each annulus is carried out the circumferencial direction discretization, use M iRepresent that the discrete of each annulus circumferencial direction count:
1. the center diameter is that the circle of d is as voxel, i.e. a M in each layer 0=1;
2. in each layer the
Figure C20051001225000053
Individual annulus, all uniform discrete becomes Θ point;
3. in each layer the
Figure C20051001225000055
Individual annulus, all uniform discrete becomes 2 Θ points;
4. in each layer the Individual annulus, all uniform discrete becomes 3 Θ points;
5. as above continue, up to N/2 annulus, uniform discrete becomes J MaxIndividual point;
6) calculate sytem matrix parameter corresponding under the aphylactic map projection angle with pixel driving method or ray-driven method;
7) use look-up table and calculate projection or back projection:
1. calculate the projection or the back projection of first angle by look-up table;
2. to other projection angle:
Figure C20051001225000064
Rearrangement is corresponding to the sytem matrix parameter of first angle;
Figure C20051001225000065
The projection or the back projection of the sytem matrix calculation of parameter correspondence that obtains by rearrangement;
8) rebuild finish after, according to display mode the image of the plain discretization of rotationally symmetric body is carried out interpolation.
In above-mentioned discretization method, the ring width that described scanning radius of turn direction discretization obtains and the sample rate of circumferencial direction are selected satisfying under the resolution requirement, and have whole symmetry at interval at all angles.
The present invention compares with prior art owing to adopted above-mentioned method step, produces following beneficial effect:
(1) avoids different projection angles are set up the needs of parameter list, thereby made the required parameter item of look-up table reduce by an order of magnitude, dwindled the limited degree of look-up table on internal memory.Suppose that object is H in the discretization number of plies of axis direction, then discrete bodies prime number order and the N in the one deck that obtains by the inventive method 2H is directly proportional, the item number of total sytem matrix parameter look-up table also with N 2H is directly proportional.The discrete bodies prime number order that traditional voxel discretization method obtains also with N 2H is directly proportional, but owing to the parameter that each projection angle is all needed to calculate separately, projection angle number and N are at the same order of magnitude, so the item number and the N of total parameter look-up table usually 3H is directly proportional.The inventive method utilizes rotational symmetry to avoid different projection angles are set up the needs of parameter list, makes the required parameter item of look-up table reduce by an order of magnitude, has reduced amount of calculation, has dwindled the limited degree of look-up table on internal memory.
(2) the inventive method concerns the sytem matrix that just can calculate from the geometrical relationship of a projection angle under all projection angles by simple geometric, can reduce the projection of ordinary meaning and the amount of calculation of back projection, thereby improves the CT reconstruction speed.
(3) the inventive method can adapt to any resolution requirement, can be by other symmetric use be further reduced the amount of tabling look-up, and the inventive method can directly apply to circular scanning and helical scanning simultaneously.
(4) because the discretization process of the inventive method is applicable to different scan modes, be independent of concrete algorithm for reconstructing, therefore can be incorporated into easily in the existing CT reconstruction software and use.
The present invention will be further described below in conjunction with the drawings and specific embodiments.
Description of drawings
Fig. 1 is a scanning radius of turn direction discretization sketch map in the inventive method;
Fig. 2 is a circumferencial direction discretization sketch map in the inventive method;
Fig. 3 is the plain discretization sketch map of rotationally symmetric body in the inventive method;
Fig. 4 is the geometrical relationship that 0 ray a when spending and b pass voxel (i, 0,0) and (i, j, 0) respectively for using the inventive method projection angle;
Fig. 5 is for using the inventive method object θ that turns clockwise StepBack ray a and b pass the geometrical relationship of voxel (i, Δ, 0) and (i, j+ Δ, 0) respectively.
The specific embodiment
Referring to Fig. 1 to Fig. 3, the inventive method step is:
1) requires calculating sampling step pitch d according to what rebuild resolution.
2) will effectively rebuild height H d along the axis direction of scanning rotating shaft and be separated into the disk that thickness is d.
3) every layer of disk scanned radius of turn direction discretization in diameter of section is the effective coverage of Nd, be separated into the annulus of N/2+1 ring width d, the center is circle, and the outermost layer annulus has only in the inner part
Figure C20051001225000081
Wide in effective reconstruction regions.
4) the voxel number in calculating every layer
Figure C20051001225000082
Θ is the projection number in the 360 degree scopes.
5) in every layer cross section, each annulus is carried out the circumferencial direction discretization, use M iRepresent that the discrete of each annulus circumferencial direction count, i is the subscript of ring:
1. the center diameter is that the circle of d is as voxel, i.e. a M in each layer 0=1.
2. in each layer the
Figure C20051001225000084
Individual annulus, all uniform discrete becomes Θ point.
3. in each layer the
Figure C20051001225000086
Individual annulus, all uniform discrete becomes 2 Θ points.
4. in each layer the
Figure C20051001225000088
Individual annulus, all uniform discrete becomes 3 Θ points.
5. as above continue, up to N/2 annulus, uniform discrete becomes J MaxIndividual point.
6) calculate the corresponding down sytem matrix parameter of 0 degree projection angle with pixel driving method or ray-driven method.
7) use look-up table and calculate projection or back projection:
1. calculate the projection or the back projection of first angle by look-up table;
2. to other projection angle:
Figure C20051001225000091
Rearrangement is corresponding to the sytem matrix parameter of first angle.Referring to Fig. 4 and Fig. 5, at the θ that turns clockwise StepBehind the projection angle, voxel (i, 0,0) and (i, j, 0) is in full accord with the geometrical relationship of ray a and b when the geometrical relationship of voxel (i, Δ, 0) and (i, j+ Δ, 0) and ray a and b and 0 angle projection angle.Therefore, the parameter to other projection angle can obtain from the result of calculation of first angle.The voxel under this different projection angle and the corresponding relation of detector are as shown in table 1, and θ wherein and b represent the ordinal number of projection angle respectively and at the position of detector index of some angles.
Figure C20051001225000092
Table 1
Figure C20051001225000093
The projection or the back projection of the sytem matrix calculation of parameter correspondence that obtains by rearrangement.
8) rebuild finish after, according to display mode the image of the plain discretization of rotationally symmetric body is carried out interpolation.
The special parameter of in the foregoing description, describing, select other parameter also to belong to the scope of protection of the invention as the case may be.

Claims (2)

1, a kind of plain discretization method of rotationally symmetric body that is used for the CT reconstruction, its key step is:
1) requires calculating sampling step pitch d according to what rebuild resolution;
2) axis direction of edge scanning rotating shaft will effectively be rebuild and highly be separated into the disk that thickness is d;
3) every layer of disk scanned radius of turn direction discretization in diameter of section is the effective coverage of Nd, be separated into the annulus of N/2+1 ring width d;
4) the voxel number in calculating every layer
Figure C2005100122500002C1
Θ is the projection number in the 360 degree scopes;
5) in every layer cross section, each annulus is carried out the circumferencial direction discretization, use M iRepresent that the discrete of each annulus circumferencial direction count:
1. the center diameter is that the circle of d is as voxel, i.e. a M in each layer 0=1;
2. in each layer the
Figure C2005100122500002C2
Individual annulus, all uniform discrete becomes Θ point;
3. in each layer the
Figure C2005100122500002C3
Individual annulus, all uniform discrete becomes 2 Θ points;
4. in each layer the Individual annulus, all uniform discrete becomes 3 Θ points;
5. as above continue, up to N/2 annulus, uniform discrete becomes J MaxIndividual point;
6) calculate sytem matrix parameter corresponding under the aphylactic map projection angle with pixel driving method or ray-driven method;
7) use look-up table and calculate projection or back projection:
1. calculate the projection or the back projection of first angle by look-up table;
2. to other projection angle:
Figure C2005100122500003C1
Rearrangement is corresponding to the sytem matrix parameter of first angle;
The projection or the back projection of the sytem matrix calculation of parameter correspondence that obtains by rearrangement;
8) rebuild finish after, according to display mode the image of the plain discretization of rotationally symmetric body is carried out interpolation.
2, the plain discretization method of rotationally symmetric body that is used for the CT reconstruction according to claim 1, it is characterized in that, the ring width that described scanning radius of turn direction discretization obtains and the sample rate of circumferencial direction are selected satisfying under the resolution requirement, and have whole symmetry at interval at all angles.
CNB2005100122508A 2005-07-22 2005-07-22 Rotary symmetric voxel discretization method for CT reestablishing Active CN100444798C (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CNB2005100122508A CN100444798C (en) 2005-07-22 2005-07-22 Rotary symmetric voxel discretization method for CT reestablishing

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CNB2005100122508A CN100444798C (en) 2005-07-22 2005-07-22 Rotary symmetric voxel discretization method for CT reestablishing

Publications (2)

Publication Number Publication Date
CN1899219A CN1899219A (en) 2007-01-24
CN100444798C true CN100444798C (en) 2008-12-24

Family

ID=37655345

Family Applications (1)

Application Number Title Priority Date Filing Date
CNB2005100122508A Active CN100444798C (en) 2005-07-22 2005-07-22 Rotary symmetric voxel discretization method for CT reestablishing

Country Status (1)

Country Link
CN (1) CN100444798C (en)

Families Citing this family (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CA2825704A1 (en) * 2011-01-21 2012-07-26 Mercury Computer Systems, Inc. Optimized implementation of back projection for computed tomography (ct)
CN103181770B (en) * 2011-12-30 2014-12-17 上海联影医疗科技有限公司 Spiral computed tomography (CT) three-dimensional image reconstruction method and minimum projection number determination method in reconstruction
FR3021796B1 (en) 2014-06-02 2016-06-24 Safran METHOD AND DEVICE FOR DIGITAL RECONSTRUCTION OF AN ELEMENTARY VOLUME REPRESENTATIVE OF A MICROSTRUCTURE OF COMPOSITE MATERIAL

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US4374419A (en) * 1979-07-06 1983-02-15 U.S. Philips Corporation Device for determining a radiation attenuation distribution in a plane of a body
CN1037421A (en) * 1988-04-13 1989-11-22 模拟公司 Backprojection apparatus and method
US5800355A (en) * 1995-09-27 1998-09-01 Kabushiki Kaisha Toshiba Image processing apparatus and method for quantitatively detecting and correcting the twist between images
JP2001292990A (en) * 2000-02-10 2001-10-23 Japan Science & Technology Corp Computerized tomographic image reconstruction system
CN1321884A (en) * 2001-06-08 2001-11-14 清华同方核技术股份有限公司 Taxonomic real-time reconstruction checking method of large-size workpiece, defect
CN1541617A (en) * 2002-10-22 2004-11-03 ��ʽ���綫֥ X-ray computed tomographic apparatus, image processing apparatus, and image processing method

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US4374419A (en) * 1979-07-06 1983-02-15 U.S. Philips Corporation Device for determining a radiation attenuation distribution in a plane of a body
CN1037421A (en) * 1988-04-13 1989-11-22 模拟公司 Backprojection apparatus and method
US5800355A (en) * 1995-09-27 1998-09-01 Kabushiki Kaisha Toshiba Image processing apparatus and method for quantitatively detecting and correcting the twist between images
JP2001292990A (en) * 2000-02-10 2001-10-23 Japan Science & Technology Corp Computerized tomographic image reconstruction system
CN1321884A (en) * 2001-06-08 2001-11-14 清华同方核技术股份有限公司 Taxonomic real-time reconstruction checking method of large-size workpiece, defect
CN1541617A (en) * 2002-10-22 2004-11-03 ��ʽ���綫֥ X-ray computed tomographic apparatus, image processing apparatus, and image processing method

Also Published As

Publication number Publication date
CN1899219A (en) 2007-01-24

Similar Documents

Publication Publication Date Title
CN104107065B (en) Optimal conversion of 3D image sets between different spaces
CN101470180B (en) The method and apparatus of distortion calibration in magnetic resonance imaging
CN100524358C (en) Method for removing improved conical bind CT ring shaped false shadow
Daube-Witherspoon et al. Application of the row action maximum likelihood algorithm with spherical basis functions to clinical PET imaging
US8116426B2 (en) Computed tomography device and method using circular-pixel position-adaptive interpolation
US20040207642A1 (en) Determining a coverage mask for a pixel
CN101095617A (en) X-ray CT device
CN100444798C (en) Rotary symmetric voxel discretization method for CT reestablishing
CN105931202A (en) Geometric correction die body correction method and system
CN107016672B (en) Reconstruction method and device of medical scanning image and medical imaging system
US9299140B2 (en) Method for reconstructing CT image data with weighted backprojection, including computation unit and CT system for the method
US7978810B2 (en) Imaging method for variable pitch spiral CT and a CT machine for carrying out the method
CN101398397A (en) CT imaging method of multitime scanning mode
CN103714578A (en) Single-layer rearrangement filtered backprojection reconstruction method aiming to half mulching helical cone beam CT
CN102456227A (en) Reconstruction method and device for CT (computerized tomography) image
EP1713392A1 (en) Computed tomography imaging with pixel staggering and focal spot modulation
CN109146987B (en) GPU-based rapid cone beam computed tomography reconstruction method
CN1816829A (en) Selection of a mipmap level
US8724876B2 (en) Method and system for substantially reducing streak artifacts in helical cone beam computer tomography (CT)
Hein et al. Feldkamp‐based cone‐beam reconstruction for gantry‐tilted helical multislice CT
CN1416781A (en) Fast iterative reconstruction method of high-resolution heat source image in positive-electron tomographic scanning
CN105973917B (en) X ray CT turntable unilateral side helical scanning single-layer rearrangement method for reconstructing twice
US20060104408A1 (en) Methods, apparatus, and software to facilitate computing the elements of a forward projection matrix
Zamyatin et al. Practical hybrid convolution algorithm for helical CT reconstruction
US5126938A (en) Computerized tomography system

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