Radial Adaptive Filter for Metal Artifact Correction

• US 20080152203A1
• Filed: 02/01/2006
• Published: 06/26/2008
• Est. Priority Date: 02/03/2005
• Status: Active Grant

• Current Assignee: Koninklijke Philips Electronics N.V. (Koninklijke Philips N.V.)
• Original Assignee: Koninklijke Philips Electronics N.V. (Koninklijke Philips N.V.)
• Inventors: Spies, Lothar, Bal, Matthieu, Celik, Hasan, Subramanyan, Krishna, Eck, Kai
• Granted Patent: US 7,991,243 B2
• Time in Patent Office: Days
• US Class Current: 382/131
• CPC Class Codes: G06T 11/008 Specific post-processing after tomographic reconstruction, e.g. voxelisation, metal artifact correction

Abstract

A diagnostic imaging system (10) corrects metal artifact streaks (38) emanating from a metal object (36) in a tomographic image (T). A first processor (40) reduces streaks (38) caused by mild artifacts by applying an adaptive filter (82). The filter (82) is perpendicularly oriented toward the center of the metal object (36). The weight of the filter (82) is a function of the local structure tensor and the vector pointing to the metal object (36). If it is determined that the strong artifacts are present in the image, a second processor (48) applies a sinogram completed image algorithm to correct for severe artifacts in the image. The sinogram completed image and adaptively filtered image are fused to a final corrected image. In the fusion process, highly corrupted tomographic regions are replaced by the result of the sinogram completed image and the remainder is replaced by the adaptively filtered image.

Claims

1. A diagnostic imaging system which corrects metal artifact streaks emanating from a high attenuating object in an uncorrected tomographic image, the system comprising:

a first vector means for determining a direction vector for each pixel, which direction vector points in the direction of the high attenuating object;

a second vector means for determining an orientation vector for each pixel, which orientation vector coincides with a steepest gradient direction; and

an adaptive filter means for adaptively filtering the uncorrected tomographic image based at least on the determined orientation of the steepest gradient vector.

2. The system as set forth in claim 1, wherein the adaptive filter means includes:

an adjustable dynamic adaptive filter which is applied at least to pixels of the tomographic image that are intersected by the streak; and

a filter adjusting means for adjusting the adaptive filter in accordance with a degree to which each filtered pixel is corrupted by the streak.

3. The system as set forth in claim 1, further including:

a means for determining a size of the steepest gradient; and

a threshold means for determining if the steepest gradient indicates a streak by at least one of;

comparing the size of the steepest gradient to a prespecified threshold value, and

comparing the direction of the steepest gradient to a direction perpendicular to the direction vector.

4. The system as set forth in claim 3, wherein the adaptive filter includes an adaptive asymmetric Gaussian filter which is adjustable to be oriented orthogonal to the vector pointing in the direction of the high attenuating object.

5. The system as set forth in claim 4, further including:

a means for adjusting the adaptive filter to have a largest width in the direction perpendicular to the direction vector and a smallest width in a direction parallel to the direction vector such that the biggest correction is performed in the direction perpendicular to the direction vector.

6. The system as set forth in claim 4, wherein the Gaussian filter is scalable in accordance with a magnitude of the steepest gradient.

7. The system as set forth in claim 1, further including:

a second correction means for reducing severe artifacts in the tomographic image.

8. The system as set forth in claim 7, wherein the second correction means applies a sinogram completion algorithm to reduce severe artifacts.

9. The system as set forth in claim 8, further including:

a means for fusing the adaptively corrected image and sinogram completed image into a 3D image representation.

10. A method of diagnostic imaging, which corrects metal artifact streaks emanating from a high attenuating object in an uncorrected tomographic image, comprising:

determining a direction vector for each pixel which direction vector points in the direction of the high attenuating object;

determining an orientation vector for each pixel which orientation vector coincides with a direction of steepest gradient; and

adaptively filtering the uncorrected tomographic image based at least on the determined orientation of the steepest gradient vector.

11. The method as set forth in claim 10, wherein the step of adaptive filtering includes:

applying an adjustable dynamic adaptive filter at least to pixels of the tomographic image that are intersected by the streak; and

adjusting the adaptive filter in accordance with a degree to which each filtered pixel is corrupted by the streak.

12. The system as set forth in claim 10, further including:

adjusting the adaptive filtering to have a largest width in the direction perpendicular to the direction vector and a smallest width in a direction parallel to the direction vector such that the biggest correction is performed in the direction perpendicular to the direction vector.

13. The method as set forth in claim 10, further including:

determining a size of the steepest gradient; and

determining whether the steepest gradient indicates a streak by at least one of;

comparing the size of the steepest gradient to a prespecified threshold value, and

comparing the direction of the steepest gradient to a direction perpendicular to the direction vector.

14. The method as set forth in claim 13, wherein the adaptive filter is an asymmetric Gaussian filter which is oriented in one of:

orthogonal to the direction of the high attenuating object, and

parallel to the direction of the steepest gradient.

15. The method as set forth in claim 14, wherein the Gaussian filter is scalable in accordance with a magnitude of the steepest gradient.

16. The method as set forth in claim 10, further including:

reducing severe artifacts in the tomographic image.

17. The method as set forth in claim 16, wherein the step of artifacts reduction includes:

applying a sinogram completion algorithm to reduce severe artifacts.

18. The method as set forth in claim 17, further including:

fusing the adaptively corrected image and sinogram completed image into a 3D image representation.

19. A method of adaptive filtering of image data comprising:

receiving image data;

reconstructing the data into a tomographic image;

comparing a value of each pixel of the tomographic image to a prespecified threshold to identify one of uncorrupted pixel, severe corrupted pixel and mildly corrupted pixel;

adapting an adaptive filter at least in accordance with mild corruption to each mildly corrupted pixel; and

applying the adaptive filter to the identified mildly corrupted pixel of the tomographic image data to correct and harmonize minor noise.

20. The method as set forth in claim 19, further including:

applying a major correction algorithm which corrects severe noise to the identified severe corrupted pixels; and

fusing the adaptively corrected image data and major corrected image data into a 3D corruption reduced image representation.