Image reconstruction method for computed tomography

• US 6,839,400 B2
• Filed: 12/06/2002
• Issued: 01/04/2005
• Est. Priority Date: 12/06/2001
• Status: Active Grant

• Current Assignee: Siemens Healthcare GMBH (Siemens AG)
• Original Assignee: Siemens AG
• Inventors: Bruder, Herbert, Flohr, Thomas, Stierstorfer, Karl
• Primary Examiner(s): Bruce, David V.
• Assistant Examiner(s): KEANEY, ELIZABETH MARIE
• Application Number:
  US10/310,927
• Publication Number:
  US 20030133533A1
• Time in Patent Office: 760 Days
• Field of Search: 378/4, 378/8, 378/15, 378/19, 378/901
• US Class Current: 378/4
• CPC Class Codes: A61B 6/027 characterised by the use of a particular data acquisition trajectory, e.g. helical or spiral
• CPC Class Codes: G06T 17/00 Three dimensional [3D] modelling, e.g. data description of 3D objects
• CPC Class Codes: Y10S 378/901 Computer tomography program or processor

Abstract

A method for computed tomography, includes, in order to scan an object to be examined with a conical beam originating from a focus and with a matrix-like detector array for detecting the beam, the focus is moved, relative to the object to be examined, on a focal path around a system axis. A detector array supplies an output data corresponding to received radiation. The output data is filtered, and the filtered output data is backprojected three-dimensionally in order to produce at least one slice of a layer of the, which has a layer thickness. The slice represents absorption values obtained from the output data of a voxel belonging to the layer for the radiation from the beam. Filtering is carried out at least also in the direction of the tangent to the focal path belonging to a respective focal position, and normalization is carried out for each voxel considered.

Claims

1. An image reconstruction method for computed tomography, comprising:

moving a focus on a focal path around a system axis relative to an object using a conical beam originating from the focus and a matrix-like detector array for detecting the beam;

supplying the detector array an output data corresponding to received radiation in the form of rays with a fan ray geometry;

filtering the output data; and

backprojecting the filtered output data three-dimensionally to produce at least one slice of a layer of the object, the slice representing absorption values obtained from a voxel associated with the layer of the object,

wherein the filtering is carried at least in a direction of a tangent to the focal path belonging to a respective focal position, and normalization is carried out for each voxel considered.

2. The method as claimed in claim 1, wherein before the filtering, the method includes converting output data in fan ray geometry in the form of rays P(α

,β

,q) into parallel data present in parallel ray geometry in the form of rays P(θ

,β

,q) and P(θ

,p,q), where

α

is a focus angle

β

is a fan angle

q is a line index of the detector system corresponding to the z coordinate,

θ

=α

+β

is a parallel fan angle

p=R_F sin(β

) is a parallel coordinate corresponding to the distance of the ray from the axis of rotation (system axis), and

R_F is the radius of a focal path.

3. The method as claimed in claim 2, further comprising forming, in the course of the backprojection for each voxel (x,y,z) for each θ

ε

[0, π

] for the rays P(θ

+kπ

, {tilde over (β

)}, q) and P(θ

+kπ

, {tilde over (p)}, q), whose projection along the system axis goes through (x,y), the sum $P_{x,y,z}<br><br>\left(\mathrm{\theta <br><br>}\right)=\underset{k}{\sum <br><br>}<br><br>\underset{q}{\sum <br><br>}<br><br>h<br><br>\left(d_{x,y,z}<br><br>\left(\mathrm{\theta <br><br>}+k<br><br>\text{\hspace{1em}<br><br>}<br><br>\mathrm{\pi <br><br>},\left\{\begin{array}{c}\tilde{P}\\ \widetilde{\mathrm{\beta <br><br>}}\end{array}\right\},q\right)\right)<br><br>P<br><br>\left(\mathrm{\theta <br><br>}+k<br><br>\text{\hspace{1em}<br><br>}<br><br>\mathrm{\pi <br><br>},\left\{\begin{array}{c}\tilde{P}\\ \widetilde{\mathrm{\beta <br><br>}}\end{array}\right\},q\right)$

where

x,y,z are coordinates of the respective voxel,

k is a whole number corresponding to a number of half-revolutions of the focus included in a reconstruction,

{tilde over (p)} are parallel coordinates of those rays whose projections along the system axis run through coordinates (x,y) of a respective voxel (x,y,z),

{tilde over (β

)} arefan angles of those rays whose projections along the system axis run through the coordinates (x,y) of a respective voxel (x,y,z), and

h is a weighting function that determines the layer thickness of the layer of the object displayed in the slice generated, and

d is a function which is equal to a distance of the respective ray from the corresponding voxel (x,y) or depends on the distance of a respective ray from the corresponding voxel (x,y).

4. The method as claimed in claim 3, further comprising normalizing and forming, during the backprojection of the parallel data, the sum to the sum H of the weights h, where $$

⁢

H=∑

 

⁢

k⁢

 

⁢

∑

 

⁢

q⁢

h⁡

(dx, 

⁢

y, 

⁢

z⁡

(θ

+k⁢

 

⁢

π

, 

⁢

{p~β

~}, 

⁢

q))$P_{x,\text{\hspace{1em}<br><br>}<br><br>y,\text{\hspace{1em}<br><br>}<br><br>z}<br><br>\left(\mathrm{\theta <br><br>}\right)=\frac{1}{H}<br><br>\underset{k}{\sum <br><br>}<br><br>\text{\hspace{1em}<br><br>}<br><br>\underset{q}{\sum <br><br>}<br><br>\text{\hspace{1em}<br><br>}<br><br>h<br><br>\left(d_{x,\text{\hspace{1em}<br><br>}<br><br>y,\text{\hspace{1em}<br><br>}<br><br>z}<br><br>\left(\mathrm{\theta <br><br>}+k<br><br>\text{\hspace{1em}<br><br>}<br><br>\mathrm{\pi <br><br>},\text{\hspace{1em}<br><br>}<br><br>\left\{\begin{array}{c}\tilde{p}\\ \widetilde{\mathrm{\beta <br><br>}}\end{array}\right\}<br><br>,\text{\hspace{1em}<br><br>}<br><br>\text{\hspace{1em}<br><br>}<br><br>q\right)\right)<br><br>P<br><br>\left(\mathrm{\theta <br><br>}+k<br><br>\text{\hspace{1em}<br><br>}<br><br>\mathrm{\pi <br><br>}<br><br>,\text{\hspace{1em}<br><br>}<br><br>\text{\hspace{1em}<br><br>}<br><br>\left\{\begin{array}{c}\tilde{p}\\ \widetilde{\mathrm{\beta <br><br>}}\end{array}\right\},\text{\hspace{1em}<br><br>}<br><br>q\right).$

5. The method as claimed claim 1, wherein the filtering is carried out in the direction of the tangent to the focal path belonging to the respective focal position.

6. The method as claimed in claim 1, wherein the focal path is a circular path.

7. The method as claimed in claim 1, wherein the focal path is a spiral path, the spiral path being generated as the focus is moved on a circular path about the system axis.

8. The method as claimed in claim 7, wherein the spiral path is further generated in accordance with a relative movement of the focus in a direction of the system axis.

9. The method as claimed claim 2, wherein the filtering is carried out in the direction of the tangent to the focal path belonging to the respective focal position.

10. The method as claimed claim 3, wherein the filtering is carried out in the direction of the tangent to the focal path belonging to the respective focal position.

11. The method as claimed claim 5, wherein the filtering is carried out in the direction of the tangent to the focal path belonging to the respective focal position.

12. The method as claimed in claim 2, wherein the focal path is a circular path.

13. The method as claimed in claim 3, wherein the focal path is a circular path.

14. The method as claimed in claim 4, wherein the focal path is a circular path.

15. The method as claimed in claim 5, wherein the focal path is a circular path.

16. An image reconstruction method for computed tomography, comprising:

obtaining output data corresponding to a scanned object;

filtering the output data; and

backprojecting the filtered output data three-dimensionally to produce at least one slice of a layer of the object, the slice representing absorption values obtained from a voxel associated with the layer of the object,

wherein the filtering is carried at least in a direction of a tangent to the focal path belonging to a respective focal position, and normalization is carried out for each voxel considered.

17. The method as claimed in claim 16, wherein before the filtering, the method includes converting the output data in the form of rays P(α

,β

,q) into parallel data present in parallel ray geometry in the form of rays P(θ

,β

,q) and P(θ

,p,q), where

α

is a focus angle

β

is a fan angle

q is a line index of the detector system corresponding to the z coordinate,

θ

=α

+β

is a parallel fan angle

p=R_F sin(β

) is a parallel coordinate corresponding to the distance of the ray from the axis of rotation (system axis), and

R_F is the radius of a focal path.