Method of reconstructing computed tomography (CT) volumes suitable for execution on commodity central processing units (CPUs) and graphics processors, and apparatus operating in accord with those methods (rotational X-ray on GPUs)

• US 7,778,392 B1
• Filed: 11/01/2005
• Issued: 08/17/2010
• Est. Priority Date: 11/02/2004
• Status: Active Grant

• Current Assignee: PME IP Pty., Ltd.
• Original Assignee: PME IP Australia Pty Ltd. (Pro Medicus Limited)
• Inventors: Goddard, Joseph, Thieret, Scott A., Berman, Ari P.
• Primary Examiner(s): Glick; Edward J
• Assistant Examiner(s): Corbett; John M
• Application Number:
  US11/264,287
• Time in Patent Office: 1,750 Days
• Field of Search: 378/4, 378/210, 378/901, 345/424, 345/426, 345/506, 345/586
• US Class Current: 378/210
• CPC Class Codes: A61B 6/027 characterised by the use of a particular data acquisition trajectory, e.g. helical or spiral
• CPC Class Codes: A61B 6/032 Transmission computed tomography [CT]
• CPC Class Codes: A61B 6/583 using calibration phantoms
• CPC Class Codes: G06T 11/006 Inverse problem, transformation from projection-space into object-space, e.g. transform methods, back-projection, algebraic methods
• CPC Class Codes: G06T 2211/421 Filtered back projection [FBP]
• CPC Class Codes: G06T 2211/428 Real-time
• CPC Class Codes: Y10S 378/901 Computer tomography program or processor

Abstract

The invention provides in one aspect methods and apparatus for use with C-arm and other CT systems, e.g., with non-rigid geometries. In such systems, by way of example, calibration can be performed to determine the exact position of the x-ray source and the exact orientation of the detector where each projection measurement is made. Next, a weighting coefficient can be determined for the voxels in each plane of a reconstruction volume at every possible projection. Finally, the order in which to process the voxels during image reconstruction can be determined. Following an actual CT scan procedure in which scans are obtained of a volume to be constructed, a system according to these and related aspects of the invention can use an optimal, pre-calculated processing method, while utilizing offsets and weighting coefficients determined during calibration, for performing backprojection image reconstruction.

Claims

1. A method for computed tomography (CT) volume reconstruction, the method comprising the steps of:

A. during a precalculation phase,

(i) determining a normalization to compensate for a non-circular orbit of a CT scanner;

(ii) determining from one or more images acquired by the CT scanner of a known volume a weighting coefficient for each of one or more voxels of that volume; and

(iii) determining an order in which to process two or more bricks of a volume to be reconstructed based at least in part on the proximity of said bricks to one another when images of the bricks are projected onto a detector plane, wherein each brick comprises voxels;

B. during a runtime phase,

(i) acquiring images from the CT scanner at one or more image locations of the volume to be reconstructed; and

(ii) performing volume reconstruction from the images of the volume to be reconstructed utilizing a volume reconstruction algorithm and the steps of the precalculation phase.

2. The method of claim 1, wherein upon executing step A(i), determining the normalization of step A(i) comprises imaging a calibration phantom at one or more image locations, where a composition of the calibration phantom is known thereby allowing for a comparison between an expected result after volume reconstruction and an actual result obtained after volume reconstruction of the calibration phantom is performed.

3. The method of claim 1, wherein upon executing step A(ii), the weighting coefficient for each voxel is calculated as a function of a distance from a radiation source of the CT scanner to a reconstruction volume of interest and an angle of the image position.

4. The method of claim 3, wherein upon executing steps A(i) and A(ii), the weighting coefficient for each voxel is further calculated as a function of a transform based on the normalization to compensate for the non-circular orbit of the non-rigid CT scanner as found in step A(i).

5. The method of claim 1, wherein upon executing step A(ii), the weighting coefficient for each voxel is determined by calculating a single plane of weighting coefficients for each image of the volume to be reconstructed.

6. The method of claim 1, wherein step A(iii) further comprises determining an order to process the voxels in each brick.

7. The method of claim 1, wherein performing volume reconstruction comprises:

dividing the volume to be reconstructed into one or more slabs; and

processing the voxels for each brick for each slab using the processing order found in step A(iii).

8. The method of claim 1, wherein upon executing either step A(i) or A(ii), the CT scanner includes a C-arm.

9. The method of claim 1, wherein one or more of the steps of the precalculation phase are performed on a central processing unit.

10. The method of claim 1, wherein said one or more of the steps of the precalculation phase are performed on a commodity graphical processing unit (GPU).

11. The method of claim 1, wherein upon executing step A(ii), the weighting coefficients are found for each voxel in a single plane in the one or more images acquired by the CT scanner of the known volume.

12. The method of claim 1, wherein upon executing step A(ii), the weighting coefficients are found for each voxel in the one or more images acquired by the CT scanner of the known volume.

13. A method for computed tomography (CT) for determining an order in which to process voxels during volume reconstruction of an object, the method comprising:

A. during a precalculation phase, executing the steps of,

(i) defining a brick size for a plurality of projection images of a volume to be reconstructed, wherein the volume to be reconstructed comprises a plurality of bricks;

(ii) defining one or more projection sets comprising a sub-set of the plurality of projection images acquired at a plurality of projection locations of the volume to be reconstructed, the projection images in each projection set utilizing an order for processing the bricks;

(iii) building a table of u-axis and v-axis coordinates at which each brick in the projection images of the volume to be reconstructed is projected onto a resampled detector plane, the u-axis and v-axis being in a plane parallel to the resampled detector plane, wherein the resampled detector plane is in a plane tangential to a resampled circle at each projection position, wherein the resampled circle is a best-fit circle representing a non-circular orbit of a CT scanner based on each projection position;

determining the order to process the bricks;

determining an order to process voxels contained in each brick; and

B. during a runtime phase, applying the processing orders determined in the precalculation phase during volume reconstruction of the volume to be reconstructed.

14. The method of claim 13, wherein the brick size defined in step A(i) is a multiple of a number of pixel shaders used to process the bricks.

15. The method of claim 13, wherein the brick size does not exceed a cache limit of each pixel shader.

16. The method of claim 13, wherein the table of u-axis and v-axis coordinates are stored in a memory of a digital data processor.

17. The method of claim 13, wherein a number of tables stored in a memory corresponds to the number of projection positions.

18. The method of claim 13, wherein defining the order to process the bricks is based at least in part on the proximity of the bricks to one another when projected onto the resampled detector plane.

19. The method of claim 13, wherein defining the order to process voxels includes selecting a primary axis to occupy consecutive locations in memory of a digital data processor.

20. The method of claim 13, wherein the precalculation phase further comprises:

(vi) repeating step A(v) for each brick of the projection images of the volume to be reconstructed.

21. A method for computed tomography (CT) for determining an order in which to process voxels during volume reconstruction of an object, the method comprising:

A. during a precalculation phase, executing the steps of,

(i) defining a brick size for a plurality of projection images of a volume to be reconstructed, wherein the volume to be reconstructed comprises a plurality of bricks;

(ii) defining one or more projection sets comprising a subset of the plurality of projection images acquired at a plurality of projection locations of the volume to be reconstructed, the plurality of projection images in each projection set utilizing an order for processing the bricks;

(iii) building a table of u-axis and v-axis coordinates at which each brick in the projection images of the volume to be reconstructed is projected onto a detector plane, the u-axis and v-axis being in a plane parallel to the detector plane, wherein the detector plane is in a plane tangential to a circle at each projection position, wherein the circle represents an orbit of a CT scanner based on each projection position;

(iv) determining the order to process the bricks based at least in part on the proximity of said bricks to one another when projected onto the detector plane;

(v) determining an order to process voxels contained in each brick; and

B. during a runtime phase, applying the processing orders determined in the precalculation phase during volume reconstruction of the volume to be reconstructed.

22. The method of claim 21, further comprising, during the precalculation phase, executing the steps of:

(vi) repeating step A(v) for each brick of the one or more projection images of the volume to be reconstructed.

23. The method of claim 21, wherein determining the order to process the voxels is based at least in part on selecting a primary axis to occupy consecutive locations in memory of a digital data processor.

24. The method of claim 21, wherein step A(ii) comprises:

defining one or more projection sets that comprise the subset of the plurality of projection images acquired at a plurality of projection locations of a volume to be reconstructed, the projection images in each projection set utilizing the same order for processing the bricks.

25. The method of claim 21, wherein the orbit of the CT scanner is a non-circular orbit and the circle is a best-fit circle representing the non-circular orbit of the CT scanner based on each projection position.

26. A method for computed tomography (CT) for determining an order in which to process the voxels during volume reconstruction of an object, the method comprising:

A. defining a brick size for a plurality of projection images of a volume to be reconstructed, wherein the volume to be reconstructed comprises a plurality of bricks;

B. defining one or more projection sets comprising a sub-set of the plurality of projection images acquired at a plurality of projection locations of the volume to be reconstructed, the projection images in each projection set utilizing an order for processing the bricks;

C. building a table of u-axis and v-axis coordinates at which each bricks in the projection images of the volume to be reconstructed is projected onto a detector plane, the u-axis and v-axis being in a plane parallel to the detector plane, wherein the detector plane is in a plane tangential to a circle at each projection position, wherein the circle represents an orbit of a CT scanner based on each projection position;

D. determining the order to process the bricks based at least in part on the proximity of said bricks to one another when projected onto the detector plane;

E. determining an order to process the voxels contained in each brick; and

F. applying the processing orders determined in steps D and E during volume reconstruction of the volume to be reconstructed.

27. The method of claim 26, further comprising repeating step E for each brick of the projection images of the volume to be reconstructed.

28. The method of claim 26, wherein determining the order to process the voxels is based at least in part on selecting a primary axis to occupy consecutive locations in memory of a digital data processor.

29. The method of claim 26, wherein steps A through E are executed during a precalculation phase.

30. The method of claim 29, wherein step F is executed during a runtime phase.

31. The method of claim 26, wherein the orbit of the CT scanner is a non-circular orbit and the circle is a best-fit circle representing the non-circular orbit of the CT scanner based on each projection position.

32. A method for computed tomography (CT) for determining an order in which to process the voxels during volume reconstruction of an object, the method comprising:

A. defining a brick size for a plurality of projection images of a volume to be reconstructed, wherein the volume to be reconstructed comprises a plurality of bricks;

B. defining one or more projection sets comprising a sub-set of the plurality of projection images acquired at a plurality of projection locations of the volume to be reconstructed, the projection images in each projection set utilizing an order for processing the bricks;

C. building a table of u-axis and v-axis coordinates at which each bricks in the projection images of the volume to be reconstructed is projected onto a detector plane, the u-axis and v-axis being in a plane parallel to the detector plane, wherein the detector plane is in a plane tangential to a circle at each projection position, wherein the circle represents an orbit of a CT scanner based on each projection position;

D. determining the order to process the bricks based at least in part on the proximity of said bricks to one another when projected onto the detector plane;

E. determining an order to process the voxels contained in each brick based at least in part on selecting a primary axis to occupy consecutive locations in memory of a digital data processor; and

F. applying the processing orders determined in steps D and E during volume reconstruction of the volume to be reconstructed.

33. The method of claim 32, further comprising:

repeating step E for each brick of the projection images of the volume to be reconstructed.

34. The method of claim 32, wherein steps A through E are executed during a precalculation phase.

35. The method of claim 34, wherein step F is executed during a runtime phase.

36. The method of claim 32, wherein the orbit of the CT scanner is a non-circular orbit and the circle is a best-fit circle representing the non-circular orbit of the CT scanner based on each projection position.