CN106447740B - Opposite parallel lines CT area-of-interest image rebuilding method - Google Patents
Opposite parallel lines CT area-of-interest image rebuilding method Download PDFInfo
- Publication number
- CN106447740B CN106447740B CN201610878379.5A CN201610878379A CN106447740B CN 106447740 B CN106447740 B CN 106447740B CN 201610878379 A CN201610878379 A CN 201610878379A CN 106447740 B CN106447740 B CN 106447740B
- Authority
- CN
- China
- Prior art keywords
- area
- image
- parallel lines
- data
- projection
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T11/00—2D [Two Dimensional] image generation
- G06T11/003—Reconstruction from projections, e.g. tomography
- G06T11/006—Inverse problem, transformation from projection-space into object-space, e.g. transform methods, back-projection, algebraic methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2210/00—Indexing scheme for image generation or computer graphics
- G06T2210/41—Medical
Landscapes
- Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Algebra (AREA)
- Mathematical Analysis (AREA)
- Mathematical Optimization (AREA)
- Mathematical Physics (AREA)
- Pure & Applied Mathematics (AREA)
- Image Processing (AREA)
- Apparatus For Radiation Diagnosis (AREA)
Abstract
The present invention relates to a kind of opposite parallel lines CT area-of-interest image rebuilding methods, comprising the following steps: S1. obtains data for projection；S2. differential back projection is carried out to data for projection；S2. Hilbert inverse transformation is carried out to the data that back projection obtains, obtains reconstruction image.The present invention can not only not be that approximate complete subject image is rebuild very under serious situation in truncation, and without gibbs artifact Exact Reconstruction region of interest area image in the truncated data that can be collected with PTCT system.In addition, the invention also provides a practical weighting functions to cut down the redundancy that general multiple linear scanning model generates.
Description
Technical field
The present invention relates to image reconstruction fields, and in particular to a kind of opposite parallel lines CT area-of-interest image reconstruction side
Method.
Background technique
The new CT system for moving in parallel (PTCT) in different directions based on x-ray source and detector of the one kind being recently proposed
System structure, and have confirmed that it has huge potentiality in inexpensive CT scanner.However, in order to optimize PTCT system, it is related
Image reconstruction should be by primary study.
Before for FBP algorithm under PTCT system, the image reconstruction [3] under complete and data without truncation is mainly handled.
However, detector can only often cover a part of object in PTCT system, this causes data to be truncated, and further results in
Exact image reconstruction is complicated, or even not can be carried out image reconstruction.
Summary of the invention
In view of this, the purpose of the present invention is to provide a kind of opposite parallel lines CT area-of-interest image reconstruction sides
Method.
The purpose of the present invention is achieved through the following technical solutions,
A kind of opposite parallel lines CT area-of-interest image rebuilding method, comprising the following steps: S1. obtains projection number
According to；S2. differential back projection is carried out to data for projection；S3. Hilbert inverse transformation is carried out to the data that back projection obtains, is weighed
Build image.
Further, the step S2. back projection step is in line style PI lineIt is upper to generate an intermediate Hilbert image letter
Number Wherein
λ show radiographic source to the vector of origin in the angle of y-axis, h is distance of the origin to x-ray source track, and SDD is ray
To the distance of detector trajectory, ψ is the angle of radiographic source track and x-axis, λ for source trackbAnd λeIt is that x-ray source track starts and ties
The angle of beam position, p refer to data for projection.
Further, in step s 2, the once linear that Hilbert inverse transformation obtains scans image obtained are as follows:
Or
Wherein L > l >=max (xe|,|xb|), k (L, l, x) is expressed as follows
Further, the l=max (| xb|,|xe|)+(2~3pixels), L=(1.1~1.3) max (| xe|,|xb|)。
Further, the true picture obtained from Hilbert image are as follows:
ε is a minimum, and value range is (10-3, 10-2)。
Further, multiple linear scan is carried out, then the reconstruction image under multiple linear scan pattern are as follows:
Or
It further, further include that step S3. cuts down the redundancy that multiple linear scanning model generates using weighting function,
The weighting function are as follows:
Further, in order to avoidDiscontinuity, then
It is a light
Sliding positive function,
Due to using the technology described above, the invention has the following advantages that
It is proposed by the present invention can not only not be in truncation approximate complete subject image is rebuild very under serious situation, and
Without gibbs artifact Exact Reconstruction region of interest area image in the truncated data that can be collected with PTCT system.
Detailed description of the invention
To make the objectives, technical solutions, and advantages of the present invention clearer, below in conjunction with attached drawing to the present invention make into
The detailed description of one step, in which:
Fig. 1 is the general linear scanning model of PTCT system；
Fig. 2 is that PTCT system data obtains model；
Fig. 3 is that assigned direction passes through data for projection of the terminal about different scanning track under parallel lines scanning；
Fig. 4 is that area-of-interest image reconstruction is truncated in detector；
Fig. 5 is Sheep-Slogan used for reconstruction figure and smiling face's figure；
As Fig. 6 uses the FBP algorithm of formula (8), formula (14), the MP-BPF algorithm of formula (26), and formula respectively
(27) MZ-BPF algorithm schemes Shepp-Logan to carry out the image reconstruction of the non-truncated data of noiseless fladellum, and the first row arrives
The third line respectively represents 1 time, 2 times and 3 times scan model；
Fig. 7 is that 2 scan models have weighting function and the reconstructed results comparison without weighting function；
Fig. 8 is MP-BPF, image profile figure of the MZ-BPF and FBP algorithm under different scan models on y=0 straight line
Gray value size；
Fig. 9 is the FBP algorithm of formula (8), formula (15), the MP-BPF algorithm of formula (27), and the MZ- of formula (28)
BPF algorithm schemes Shepp-Logan to carry out the image reconstruction of the non-truncated data of noiseless fladellum；
Figure 10 is the FBP algorithm using formula (8), formula (14), the MP-BPF algorithm of formula (26) and formula (27)
MZ-BPF algorithm schemes Shepp-Logan to carry out the image reconstruction of noiseless fladellum truncated data, the first row to the third line point
1 time, 2 times and 3 times scan model is not represented；
Figure 11 is area-of-interest image reconstruction；
Figure 12 is the sectional view in the direction y=0 in Figure 10；
Figure 13 is that the image reconstruction for having noise fladellum truncated data is；
Figure 14 is sectional view of the Figure 13 along the direction y=0.
Specific embodiment
Below in conjunction with attached drawing, a preferred embodiment of the present invention will be described in detail.
PTCT model is looked back first.In PTCT, target hovering, x-ray source and detector move in opposite direction, such as scheme
Shown in 1.Now, using target's center as the origin of cartesian coordinate system.Radiographic source track is expressed as follows
λ show radiographic source to the vector of origin in the angle of y-axis, h is distance of the origin to x-ray source track, and SDD is ray
Distance of the source track to detector trajectory.ψ is the angle of radiographic source track and x-axis.Under this work, it is assumed that image function
Now, one is described using ray source point as the moving coordinate system in the center of circle.In this system, two units to
It measures as follows
In fact, using flat panel detector in PTCT system.However, only considering fan-shaped beam scanning herein.On detector
Any parameter all indicated with t.Projection p (t, λ, ψ) is to pass through the point and detector member on radiographic source track along specified X-ray
The line integral of plain t point.Therefore it is expressed as follows in the image Function Projective p (t, λ, ψ) of t point
The unit vector for indicating the point direction from ray source point to image, is expressed as follows
According to above-mentioned equation,
λbAnd λeIt is the angle of X-ray track beginning and end position.It is reconstruction point to X ray source scanning rail
The distance of mark.G (t) is ramp filter.
For multiple opposite linear scan pattern, equation is amended as follows
According to classical BPF algorithm, objective function is made of the string for scanning track, and object can be obtained by these strings.
Classic BP F algorithm phase however for PTCT system, because all strings are covered on identical line segment, with general scanning track
Seemingly, it is believed that these special strings are linear type PI line, and the crosspoint with line style PI line and objective function is support
Section.
The back projection step of BPF algorithm is in line style PI lineIt is upper to generate an intermediate Hilbert image function
In order to simplify above-mentioned equation, the differential G (t, λ, ψ) of data for projection is expressed as follows
It is obtained by equation (5)
By equation (11), (12) bring equation (10) into
By equation (1), (8), (13) bring equation (9) into
B. Hilbert inverse transformation
F (x) is finite interval [xb,xe] on smooth function.It is its one-dimensional Hilbert transform, Xi Er
Bert is converted to as follows
Pv is the Cauchy's principal value of integral.Because target is in finite interval [xb,xe], true picture is able to use limited Xi Er
Bert transformation equation obtains.In the present invention, two limited Hilbert inverse transformation formula are used, respectively
Wherein L > l >=max (| xe|,|xb|), k (L, l, x) is expressed as follows
Wherein f (ψ, λb,λe, x) and it is f (x) in once linear scanning substantially image obtained, it is not theoretically accurately to scheme
Picture.In equation (18), the parameter sensitive to true picture is L and l.In practice, one can be selected to fit under a range
When value l=max (| xb|,|xe|)+(2~3pixels), L=(1.1~1.3) max (| xe|,|xb|)
In fact, being able to use these equations obtains true picture from Hilbert image.However, these equations have one
A singular point, in order to avoid the singular point of integral, equation can make following modification
ε is a minimum, and value range is (10-3, 10-2)。
It is based on acquisition as a result, under multiple linear scan patternIt can be reconstructed by BPF algorithm
Two formula are respectively designated as MP-BPF algorithm and MZ-BPF algorithm above.
Now, it is assumed that have n times linear scanning track.For convenience, only consider i-th and jth time linear scanning track,
As shown in Figure 3.It is defined as follows.It is a weight, in order to calculate in multiple linear scan mould
The redundancy of the data set obtained in type.Obviously, when ray passes through under different scanning tracksTerminal fan beam projections meeting
Duplicate information is provided for the reconstruction of objective function.
p(ti,λi,ψi)=p (tj,λj,ψj)s.t 1≤i,j≤N (23)
WeightFor handling redundancy.
It isThe number of hits of line and all linear scanning paths,For scanning
Fixed straight line on track by crosspoint is Φ (ψi,λi),Therefore, Weighting Functions Definitions are as follows:
SelectionAs weighting function.However, in fact, in order to avoid
Discontinuity, take following methods.
It is a smooth positive function, is denoted herein as
Finally, realizing image reconstruction with MP-BPF algorithm and MZ-BPF algorithm respectively.
Generally speaking, a series of general BPF algorithm of the reconstruction image from fladellum parallel lines scan datas is obtained.
In discussed above, such a situation is only only accounted for, is exactly that whole object can be covered (nothing section by detector
Disconnected projection).It is however quite easy to encounter such case.In PTCT system, target object cannot be completely covered by the visual field.(projection
It is truncation), as shown in Figure 5.
Linear type PI line is by parameter group (ψi,λb,λe) indicate, as formula (26) and (27), Yao Chongjian area-of-interest figure
As it is only necessary to know in [xb,xe] and [- L, L] on back projection's image.Although anti-throwing cannot be obtained from single pass model
Shadow image, but the reconstruction that complete image is realized in projection completely can be obtained by increasing the method for scan lines.Such as two
It is secondary and scan model if extending this conclusion uses the necessary and sufficient condition of BPF algorithm Exact Reconstruction image three times
Are as follows: any point in principal function will be expressed the scan angle at least needing 180 degree.If it is complete that area-of-interest is scanned line segment
It illuminates, BPF algorithm proposed by the present invention can be used and realize area-of-interest image reconstruction, and do not have gibbs artifact.
MP-BPF the and MZ-BPF algorithm proposed for fladellum PTCT system passes through an improved Shepp-Logan figure
Its validity is had found with smiling face's graph evaluation, as shown in Figure 6.It is exactly by the region that red squares cover in smiling face's figure
Area-of-interest, it is selected to table of the test MP-BPF and MZ-BPF algorithm under truncated projection compared with FBP algorithm
It is existing.The pixel of two images is all that the area size of 256 × 256 their coverings is 1 × 1mm2.For 1 time other, 2 times, 3
The parameter of secondary scanning is as shown in Table 1.It is produced under 1 time, 2 times, 3 scan patterns respectively by the detector array of 1000mm long
Raw non-truncated PTCT data.The data of truncation are obtained by the detector array of 350mm long.Nothing is added to also by Gaussian noise to make an uproar
The data with noise are generated in the data of sound.In order to prove significantly low contrast district, the maximum value of noise free data is surveyed in choosing
0.37% standard deviation as Gaussian noise.By BPF algorithm proposed in this paper and FBP algorithm to non-truncated and truncated data point
Full images reconstruction is not done and area-of-interest image reconstruction compares.
1 simulation parameter of table
As Fig. 6 uses the FBP algorithm of formula (8), formula (14), the MP-BPF algorithm and formula of formula (26) respectively
(27) MZ-BPF algorithm schemes Shepp-Logan to carry out the image reconstruction of the non-truncated data of noiseless fladellum.The first row arrives
The third line respectively represents 1 time, 2 times and 3 times scan model.In order to confirm the validity of weighting function proposed in this paper, divide in Fig. 7
Not giving 2 scan models has weighting function and the reconstructed results comparison without weighting function.Fig. 8 gives MP-BPF, MZ-
The gray value size of image profile figure of the BPF and FBP algorithm under different scan models on y=0 straight line.In order to do distinctness
Comparison, really the sectional view gray value of books is also presented in Fig. 8.The quality of reconstruction image for further evaluation arranges in table 2
The mean square error of each algorithm is gone out.
2 PTCT system of table uses the mean square error of FBP and BPF algorithm
From the above analysis, it will be apparent that it can be scanned by 2 times and three times and obtain accurate reconstruction image, 3 scanning
Picture quality be better than 2 times scanning picture qualities.The main reason for causing this result is the projection collected by 3 scanning
Data scan 360 degree around object as circular scan model, and weighting function is a simple constant 1/2 here.From another
One angle sees that 2 scan models are considered as a typical short scanning, therefore weighting function is difficult to determine.But herein
The weighting function of proposition can handle these redundancies to a certain extent, as shown in Figure 7.In addition, in scan model three times
Under, for no truncated projection data BPF algorithm quality less than FBP algorithm because using BPF algorithm can to image carry out face
Remote resetting, this can reduce picture quality.Finally, result proves MP-BPF algorithm and MZ-BPF algorithm in terms of inhibiting picture noise
There is good performance.
The area-of-interest that BPF algorithm proposed in this paper carrys out reconstruction image is applied to smiling face's picture.In this part,
It needs to observe complete approximate image to obtain, as shown in Figure 10, uses the FBP algorithm of formula (8), formula (14), formula respectively
(26) the MZ-BPF algorithm of MP-BPF algorithm and formula (27) schemes Shepp-Logan to carry out noiseless fladellum truncation number
According to image reconstruction.The first row respectively represents 1 time, 2 times and 3 times scan model to the third line.
As shown in the figure, it can be seen that the reconstruction image of no gibbs artifact can be obtained using BPF algorithm, compare FBP algorithm
Just there is gibbs artifact.In addition, BPF algorithm can be rebuild beyond area-of-interest model if projection is not seriously to be truncated very much
The complete coarse image enclosed.
In order to further study performance of the BPF algorithm proposed in this paper in terms of area-of-interest image reconstruction.Such as Figure 11
It respectively shows and obtains region of interest area image from Figure 10.In order to compare the quality of area-of-interest image reconstruction, also scheming
The sectional view of the position y=0 in Figure 11 is provided in 12.
It can be seen that in Figure 11 with red squares mark gibbs artifact be to carry out region of interest using FBP algorithm
What domain generated when rebuilding.Compare, using BPF algorithm but can not no gibbs artifact Exact Reconstruction area-of-interest.It is swept at 1 time
It retouches under model, because the loss of projection, three kinds of algorithms cannot all rebuild region of interest area image.The redundancy under 2 scan models
Information is a problem, in addition can be obtained under 3 scan patterns using BPF algorithm more complete under than 2 times scan patterns
Area-of-interest reconstruction image.
In order to further explore the BPF algorithm of PTCT system proposed in this paper to the ability inhibited in picture noise.Such as figure
13 use the FBP algorithm of formula (8), formula (14), the MP-BPF algorithm of formula (26) and the MZ-BPF algorithm pair of formula (27)
Shepp-Logan figure carries out the area-of-interest image reconstruction for having noise fladellum truncated data.Figure 14 is cuing open for the direction y=0
Face figure.These figures prove that algorithm proposed in this paper has good performance in terms of inhibiting picture noise.
Finally, it is stated that preferred embodiment above is only used to illustrate the technical scheme of the present invention and not to limit it, although logical
It crosses above preferred embodiment the present invention is described in detail, however, those skilled in the art should understand that, can be
Various changes are made to it in form and in details, without departing from claims of the present invention limited range.
Claims (7)
1. a kind of opposite parallel lines CT area-of-interest image rebuilding method, it is characterised in that: the following steps are included:
S1. data for projection is obtained；
S2. differential back projection is carried out to data for projection；
S3. Hilbert inverse transformation is carried out to the data that back projection obtains, obtains reconstruction image；
The back projection step of step S2 is in line style PI lineIt is upper to generate an intermediate Hilbert image function
Wherein
λ indicates radiographic source to the vector of origin and the angle of y-axis, and h is distance of the origin to x-ray source track, and SDD is radiographic source
To the distance of detector trajectory, ψ is the angle of radiographic source track and x-axis, λ for trackbAnd λeIt is x-ray source track beginning and end
The angle of position, p refer to data for projection.
2. opposite parallel lines CT area-of-interest image rebuilding method according to claim 1, it is characterised in that: in step
In rapid S3, the once linear that Hilbert inverse transformation obtains scans image obtained are as follows:
Or
Wherein L > l >=max (| xe|,|xb|), k (L, l, x) is expressed as follows
Wherein, L, l indicate length, they refer to that linear PI line runs through the compact schemes region of object.
3. opposite parallel lines CT area-of-interest image rebuilding method according to claim 2, it is characterised in that: described
L=max (| xb|,|xe|)+(2~3pixels), L=(1.1~1.3) max (| xe|,|xb|)。
4. opposite parallel lines CT area-of-interest image rebuilding method according to claim 3, it is characterised in that: from uncommon
The true picture obtained in your Bert image are as follows:
ε is a minimum, and value range is (10-3, 10-2)。
5. opposite parallel lines CT area-of-interest image rebuilding method according to claim 4, it is characterised in that: carry out
Multiple linear scan, the then reconstruction image under multiple linear scan pattern are as follows:
Or
N indicates n times linear scanning track.
6. opposite parallel lines CT area-of-interest image rebuilding method according to claim 5, it is characterised in that: also wrap
It includes step S3. and cuts down the redundancy that multiple linear scanning model generates using weighting function, the weighting function are as follows:
7. opposite parallel lines CT area-of-interest image rebuilding method according to claim 6, it is characterised in that: in order to
It avoidsDiscontinuity, then It is a smooth positive letter
Number,For describing the weighting function of any point in all directions in supporting zone,
Component of the parameter statement weight component of front two from radiographic source position.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610878379.5A CN106447740B (en) | 2016-10-08 | 2016-10-08 | Opposite parallel lines CT area-of-interest image rebuilding method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610878379.5A CN106447740B (en) | 2016-10-08 | 2016-10-08 | Opposite parallel lines CT area-of-interest image rebuilding method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106447740A CN106447740A (en) | 2017-02-22 |
| CN106447740B true CN106447740B (en) | 2019-04-02 |
Family
ID=58172213
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610878379.5A Active CN106447740B (en) | 2016-10-08 | 2016-10-08 | Opposite parallel lines CT area-of-interest image rebuilding method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106447740B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114708344B (en) * | 2022-03-14 | 2024-06-25 | 北京理工大学 | CT image local reconstruction system and method based on field theory |
| CN115859405B (en) * | 2023-03-02 | 2023-05-12 | 青岛昊宇重工有限公司 | Self-supporting steel chimney design data enhancement method |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104809750A (en) * | 2015-05-04 | 2015-07-29 | 重庆大学 | Linear scanning CT system and image reconstructing method |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US9600910B2 (en) * | 2014-01-08 | 2017-03-21 | Rensselaer Polytechnic Institute | Attenuation map reconstruction from TOF PET data |
-
2016
- 2016-10-08 CN CN201610878379.5A patent/CN106447740B/en active Active
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104809750A (en) * | 2015-05-04 | 2015-07-29 | 重庆大学 | Linear scanning CT system and image reconstructing method |
Non-Patent Citations (3)
| Title |
|---|
| Reconstruction From Uniformly Attenuated SPECT Projection Data Using the DBH Method;Qiu Huang et al.;《IEEE Transactions on Medical Imaging》;20090131;第28卷(第1期);第17-18页 |
| 用反投影滤波算法实现CT图像的ROI重建;洪贤勇 等;《电视技术》;20140415;第38卷(第7期);第29-32页 |
| 相对平行直线扫描CT滤波反投影图像重建;伍伟文 等;《光学学报》;20160930;第36卷(第9期);第1-11页 |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106447740A (en) | 2017-02-22 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Shukla et al. | Sampling schemes for multidimensional signals with finite rate of innovation | |
| RU2469298C2 (en) | Image reconstruction method using three-dimensional x-ray photography | |
| CN106447740B (en) | Opposite parallel lines CT area-of-interest image rebuilding method | |
| CN106056645B (en) | CT image translation motion artifact correction method based on frequency-domain analysis | |
| WO2003027954A2 (en) | Versatile cone-beam imaging apparatus and method | |
| CN109118544A (en) | Synthetic aperture imaging method based on perspective transform | |
| CN109859105A (en) | A kind of printenv image nature joining method | |
| CN111265231A (en) | Distributed light source CT image reconstruction method and system | |
| CN104735438A (en) | Multi-viewpoint data collection method based on compressive sensing | |
| AU2013273647B2 (en) | Improvements in and relating to ophthalmoscopes | |
| Li et al. | Invertible paradoxic loop structures for transformable design | |
| Stein | Geometric and photometric constraints: Motion and structure from three views | |
| Yu et al. | A geometric calibration approach for an industrial cone-beam CT system based on a low-rank phantom | |
| Valdés et al. | Camera autocalibration and the calibration pencil | |
| CN109685752A (en) | A kind of multiple dimensioned Shearlet area image method for amalgamation processing decomposed based on block | |
| US12131411B2 (en) | Apparatus and method of producing a tomogram | |
| CN110428458A (en) | Depth information measurement method based on the intensive shape coding of single frames | |
| US20220343568A1 (en) | Apparatus and method of producing a tomogram | |
| Kumar et al. | Estimation of planar angles from non-orthogonal imaging | |
| CN109003256A (en) | A kind of multi-focus image fusion quality evaluating method indicated based on joint sparse | |
| Ambartsoumian et al. | Tomographic reconstruction of nodular images from incomplete data | |
| Hou et al. | Parallel-beam ct reconstruction based on mojette transform and compressed sensing | |
| Grigoryan et al. | Novel tensor transform-based method of image reconstruction from limited-angle projection data | |
| Chen et al. | FBP-type CT reconstruction algorithms for triple-source circular trajectory with different scanning radii | |
| Heimann et al. | Jointly Resampling and Reconstructing Corrupted Images for Image Classification using Frequency-Selective Mesh-to-Grid Resampling |
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 |