Evaluating the reliability of a CT volume image

Patent metadata

Field	Value
Publication number	US-11357465-B2
Application number	US-201916598961-A
Country	US
Kind code	B2
Filing date	Oct 10, 2019
Priority date	Oct 11, 2018
Publication date	Jun 14, 2022
Grant date	Jun 14, 2022

Abstract

Evaluating the reliability of a computed tomography (CT) volume image includes acquiring a first CT volume image and a modified CT volume image that are reconstructed from scanned projection images. From the first CT volume image and the modified CT volume image, digitally reconstructed X-ray images are then calculated. A respective similarity with a corresponding one of the scanned projection images is then determined. Based on a comparison of these similarities with one another, the reliability of the CT volume images is then evaluated.

First claim

The invention claimed is:

1. A method for evaluating a reliability of computed tomography (CT) volume images of an examination object, the method comprising:
acquiring a computationally reconstructed first CT volume image from scanned projection images of the examination object;
acquiring a modified CT volume image of the examination object for image artifact reduction using an image enhancement method, the image enhancement method using a first image enhancement type;
determining at least one projection direction for which a simulated radiation beam penetrates at least one voxel of the modified CT volume image, the at least one voxel being modified in comparison with the computationally reconstructed first CT volume image;
calculating, for the at least one determined projection direction, at least one digitally reconstructed X-ray image, respectively, from the computationally reconstructed first CT volume image and from the modified CT volume image;
determining a similarity, in each case, between the at least one digitally reconstructed X-ray image calculated from the computationally reconstructed first CT volume image and the respective corresponding scanned projection image, and between the at least one digitally reconstructed X-ray image calculated from the modified CT volume image and the respective corresponding scanned projection image;
comparing, for evaluating the reliability of the CT volume images, the determined similarities with one another; and
automatically using or proposing a method for image optimization different than the image enhancement method when the reliability of the modified CT volume image is smaller than a pre-determined threshold value, the method for image optimization using a second image enhancement type, the second image enhancement type being different than the first image enhancement type.

2. The method of claim 1 , wherein the computationally reconstructed first CT volume image is already reconstructed as a CT volume image adapted through estimates based on prior knowledge for an improved image quality, and the modified CT volume image is generated by local modifications from the computationally reconstructed first CT volume image.

3. The method of claim 1 , wherein acquiring the computationally reconstructed first CT volume image comprises reconstructing the computationally reconstructed first CT volume image using only information contained directly in the scanned projection images and no estimates extending therebeyond, and
wherein acquiring the modified CT volume image comprises generating the modified CT volume image using the estimates.

4. The method of claim 1 , wherein the determined similarities are characterized by a likelihood function,
wherein the method further comprises calculating, for evaluation of the reliability using the likelihood function, in each case, a likelihood for the computationally reconstructed first CT volume and a likelihood for the modified CT volume, and
wherein the likelihood for a CT volume image is a maximum when a digitally reconstructed X-ray image calculated therefrom and the corresponding scanned projection image are identical.

5. The method of claim 4 , further comprising:
determining a noise of the scanned projection image; and
evaluating the reliability depending on whether a difference between the likelihood calculated for the computationally reconstructed first CT volume image and the likelihood calculated for the modified CT volume image is larger or smaller than the determined noise.

6. The method of claim 1 , further comprising generating, for the modified CT volume image, the computationally reconstructed first CT volume image, or the modified CT volume image and the computationally reconstructed first CT volume image, a confidence map that shows at least for each voxel modified by the image enhancement method, a corresponding reliability.

7. The method of claim 6 , wherein the reliabilities entered in the confidence map are color-coded according to a pre-defined scheme depending upon sizes.

8. The method of claim 1 , further comprising automatically issuing an indication to a user when the reliability is smaller than a pre-determined threshold value.

9. The method of claim 1 , wherein automatically using or proposing the method for image optimization using the second image enhancement type when the reliability of the modified CT volume image is smaller than the pre-determined threshold value comprises automatically using or proposing one of a method based on Baysean statistics and a method based on deep learning when the reliability of the modified CT volume image is smaller than the pre-determined threshold value, the first image enhancement type used by the image enhancement method being the other of the method based on Baysean statistics and the method based on deep learning.

10. In a non-transitory computer-readable storage medium that stores instructions executable by one or more processors to evaluate a reliability of computed tomography (CT) volume images of an examination object, the instructions comprising:
acquiring a computationally reconstructed first CT volume image from scanned projection images of the examination object;
acquiring a modified CT volume image of the examination object for image artifact reduction using an image enhancement method, the image enhancement method using a first image enhancement type;
determining at least one projection direction for which a simulated radiation beam penetrates at least one voxel of the modified CT volume image, the at least one voxel being modified in comparison with the computationally reconstructed first CT volume image;
calculating, for the at least one determined projection direction, at least one digitally reconstructed X-ray image, respectively, from the computationally reconstructed first CT volume image and from the modified CT volume image;
determining a similarity, in each case, between the at least one digitally reconstructed X-ray image calculated from the computationally reconstructed first CT volume image and the respective corresponding scanned projection image, and between the at least one digitally reconstructed X-ray image calculated from the modified CT volume image and the respective corresponding scanned projection image;
comparing, for evaluating the reliability of the CT volume images, the determined similarities with one another; and
automatically using or proposing a method for image optimization different than the image enhancement method when the reliability of the modified CT volume image is smaller than a pre-determined threshold value, the method for image optimization using a second image enhancement type, the second image enhancement type being different than the first image enhancement type.

11. The non-transitory computer-readable storage medium of claim 10 , wherein the computationally reconstructed first CT volume image is already reconstructed as a CT volume image adapted through estimates based on prior knowledge for an improved image quality, and the modified CT volume image is generated by local modifications from the computationally reconstructed first CT volume image.

12. The non-transitory computer-readable storage medium of claim 10 , wherein acquiring the computationally reconstructed first CT volume image comprises reconstructing the computationally reconstructed first CT volume image using only information contained directly in the scanned projection images and no estimates extending therebeyond, and
wherein acquiring the modified CT volume image comprises generating the modified CT volume image using the estimates.

13. The non-transitory computer-readab

Assignees

• Siemens Healthcare Gmbh

Inventors

• Fischer Peter
• Gemmel Alexander
• Kleinszig Gerhard
• Kreher Björn
• Kunze Holger
• Magaraggia Jessica
• Schneider Stefan
• Weiten Markus

Classifications

• G06T12/20

  Inverse problem, i.e. transformations from projection space into object space · CPC title

• G06T12/30 Primary

  Image post-processing, e.g. metal artefact correction · CPC title

• G06T2207/30168

  Image quality inspection · CPC title

• G06T2207/10081

  Computed x-ray tomography [CT] · CPC title

• A61B6/5258 Primary

  involving detection or reduction of artifacts or noise · CPC title