SCATTER CORRECTION METHODS

• US 20120263360A1
• Filed: 04/15/2011
• Published: 10/18/2012
• Est. Priority Date: 04/15/2011
• Status: Active Grant

• Current Assignee: Georgia Tech Research Corporation (University System of Georgia)
• Original Assignee: Georgia Tech Research Corporation (University System of Georgia)
• Inventors: ZHU, LEI, Niu, Tianye
• Granted Patent: US 8,705,827 B2
• Time in Patent Office: Days
• US Class Current: 382/131
• CPC Class Codes: G06T 11/005 Specific pre-processing for tomographic reconstruction, e.g. calibration, source positioning, rebinning, scatter correction, retrospective gating
• CPC Class Codes: G06T 2207/10081 Computed x-ray tomography [CT]
• CPC Class Codes: G06T 2207/20224 Image subtraction
• CPC Class Codes: G06T 5/50 using two or more images, e.g. averaging or subtraction
• CPC Class Codes: G06T 5/70 Denoising; Smoothing

Abstract

Described herein are improved methods for correcting cone beam computed tomography signals to reduce scatter contamination contained therein. Generally, the improved methods involve generating a plurality of two-dimensional projection images of a subject from a three-dimensional multi-detector computed tomography image of the subject. This is followed by subtracting the plurality of two-dimensional projection images from a plurality of two-dimensional cone beam projection images of the subject to produce a plurality of two-dimensional estimated error projections that comprise an estimated error in the plurality of two-dimensional cone beam projection images. The plurality of two-dimensional estimated error projection images are subtracted from the plurality of two-dimensional cone beam projection images to generate a plurality of two-dimensional corrected cone beam projection images. A three-dimensional corrected cone beam computed tomography image of the subject is then constructed from the plurality of two-dimensional corrected cone beam projection images.

Claims

1. A correction method for reducing error in cone beam computed tomography images, the method comprising:

generating a plurality of two-dimensional projection images of a subject from a three-dimensional multi-detector computed tomography image of the subject, which is spatially registered with a three-dimensional cone beam computed tomography image of the subject;

subtracting the plurality of two-dimensional projection images from a plurality of two-dimensional cone beam projection images of the subject, resulting in a plurality of two-dimensional estimated error projections, which comprise an estimated error in the plurality of two-dimensional cone beam projection images;

subtracting the plurality of two-dimensional estimated error projection images from the plurality of two-dimensional cone beam projection images to generate a plurality of two-dimensional corrected cone beam projection images; and

constructing a three-dimensional corrected cone beam computed tomography image of the subject from the plurality of two-dimensional corrected cone beam projection images.

2. The method of claim 1, wherein the estimated error in the plurality of two-dimensional cone beam projection images comprises low-frequency scattered radiation signals, beam hardening effects, detector lag, and/or detector nonlinear gain.

3. The method of claim 1, further comprising smoothing the plurality of two-dimensional estimated error projections before generating the plurality of two-dimensional corrected cone beam projection images, wherein the smoothing comprises suppressing any boundary discrepancies in the plurality of two-dimensional estimated error projections.

4. The method of claim 3, wherein the smoothing further comprises reducing a high-frequency component of the estimated error.

5. The method of claim 4, wherein the high-frequency component of the estimated error comprises error in spatial registration resulting from a deformation or change in the subject.

6. The method of claim 1, wherein the three-dimensional cone beam computed tomography image of the subject is formed from the plurality of two-dimensional cone beam projection images of the subject using a Feldkamp, Davis, and Kress algorithm.

7. The method of claim 1, further comprising:

converting the three-dimensional multi-detector computed tomography image of the subject from Hounsfield Units to linear attenuation coefficients; and

spatially registering the converted three-dimensional multi-detector computed tomography image of the subject before generating the plurality of two-dimensional multi-detector projection images of the subject.

8. The method of claim 7, wherein the spatially registering comprises:

placing a data set from the converted three-dimensional multi-detector computed tomography image of the subject in a same coordinate system as a data set from the three-dimensional cone beam computed tomography image of the subject; and

aligning a center of mass of the converted three-dimensional multi-detector computed tomography image of the subject with a center of mass of the three-dimensional cone beam computed tomography image of the subject.

9. The method of claim 8, wherein the spatially registering further comprises:

calculating a square pixel-wise difference between a volume of the converted three-dimensional multi-detector computed tomography image of the subject with a volume of the three-dimensional cone beam computed tomography image of the subject;

applying a gradient descent search algorithm to the square pixel-wise difference to determine an amount of rotation and offset for the converted three-dimensional multi-detector computed tomography image of the subject; and

rotating and offsetting the converted three-dimensional multi-detector computed tomography image of the subject by the amount determined.

10. The method of claim 9, wherein the spatially registering further comprises:

repeating the calculating, applying, and rotating and offsetting steps of claim 8 until the amount determined is below a threshold amount.

11. The method of claim 1, wherein the generating the plurality of two-dimensional multi-detector projection images of the subject comprises using Siddon'"'"' s algorithm and Beer'"'"'s Law.

12. The method of claim 1, wherein the three-dimensional corrected cone beam computed tomography image of the subject is constructed from the plurality of two-dimensional corrected cone beam projection images of the subject using a Feldkamp, Davis, and Kress algorithm.

13. The method of claim 1, further comprising using the three-dimensional corrected cone beam computed tomography image of the subject as an image for image-guided radiation therapy.

14. A correction method for reducing error in cone beam computed tomography images, the method comprising:

generating a plurality of two-dimensional projection images of a subject from a cone beam computed tomography scan of the subject;

constructing a three-dimensional cone beam computed tomography image of the subject from the plurality of two-dimensional cone beam projection images;

spatially registering a three-dimensional multi-detector computed tomography image of the subject with the three-dimensional cone beam computed tomography image of the subject;

generating a plurality of two-dimensional projection images of the subject from the three-dimensional multi-detector computed tomography image of the subject;

subtracting the plurality of two-dimensional multi-detector projection images from the plurality of two-dimensional cone beam projection images to generate a plurality of two-dimensional estimated error projections, wherein the plurality of two-dimensional estimated error projections comprise an estimated error in the plurality of two-dimensional cone beam projection images;

subtracting the plurality of two-dimensional estimated error projection images from the plurality of two-dimensional cone beam projection images to generate a plurality of two-dimensional corrected cone beam projection images; and

constructing a three-dimensional corrected cone beam computed tomography image of the subject from the plurality of two-dimensional corrected cone beam projection images.

15. The method of claim 14, further comprising:

smoothing the plurality of two-dimensional estimated error projections before generating the plurality of two-dimensional corrected cone beam projection images;

wherein the smoothing comprises;

suppressing any boundary discrepancies in the plurality of two-dimensional estimated error projections; and

reducing a high-frequency component of the estimated error, wherein the high-frequency component of the estimated error comprises error in spatial registration resulting from a deformation or change in the subject.

16. The method of claim 14, wherein the three-dimensional cone beam computed tomography image of the subject is constructed from the plurality of two-dimensional cone beam projection images of the subject using a Feldkamp, Davis, and Kress algorithm and/or wherein the three-dimensional corrected cone beam computed tomography image of the subject is constructed from the plurality of two-dimensional corrected cone beam projection images of the subject using a Feldkamp, Davis, and Kress algorithm.

17. The method of claim 14, further comprising converting the three-dimensional multi-detector computed tomography image of the subject from Hounsfield Units to linear attenuation coefficients before spatially registering the three-dimensional multi-detector computed tomography image of the subject with the three-dimensional cone beam computed tomography image of the subject.

18. The method of claim 14, wherein spatially registering the three-dimensional multi-detector computed tomography image of the subject with the three-dimensional cone beam computed tomography image of the subject comprises:

placing a data set from the three-dimensional multi-detector computed tomography image of the subject in a same coordinate system as a data set from the three-dimensional cone beam computed tomography image of the subject;

aligning a center of mass of the three-dimensional multi-detector computed tomography image of the subject with a center of mass of the three-dimensional cone beam computed tomography image of the subject;

calculating a square pixel-wise difference between a volume of the three-dimensional multi-detector computed tomography image of the subject with a volume of the three-dimensional cone beam computed tomography image of the subject;

applying a gradient descent search algorithm to the square pixel-wise difference to determine an amount of rotation and offset for the three-dimensional multi-detector computed tomography image of the subject;

rotating and offsetting the three-dimensional multi-detector computed tomography image of the subject by the amount determined; and

repeating the calculating, applying, and rotating and offsetting until the amount determined is below a threshold amount.

19. The method of claim 14, wherein the generating the plurality of two-dimensional multi-detector projection images of the subject from the three-dimensional multi-detector computed tomography image of the subject comprises using Siddon'"'"'s algorithm and Beer'"'"'s Law.

20. The method of claim 14, further comprising using the three-dimensional corrected cone beam computed tomography image of the subject as an image for image-guided radiation therapy.