Metal artifacts reduction for cone beam CT using image stacking

1. A method for processing volume image data of a subject, the method executed at least in part on a computer, comprising:

obtaining two-dimensional radiographic images of the subject on an image detector, where each of the two-dimensional radiographic images are obtained with the image detector and a radiation source at a different scan angle relative to the subject;

stacking image data from a first plurality of the two-dimensional radiographic images in an image stack so that corresponding pixel data from the detector is in register using a characteristic sinusoidal pattern for each of the first plurality of the two-dimensional radiographic images in the image stack;

segmenting pixels that represent metal objects from said each of the first plurality of the two-dimensional radiographic images of the image stack by a repeated process of obtaining an image stack slice from the image stack in a direction that is orthogonal to the obtained two-dimensional radiographic images and processing the obtained image stack slice;

replacing data for at least some of the segmented pixels to generate a second plurality of modified two-dimensional radiographic images with a reduced amount of the pixels that represent metal objects;

combining the second plurality of modified two-dimensional radiographic images with the first plurality of two-dimensional radiographic images of the subject to form a third plurality of corrected two-dimensional radiographic images;

reconstructing a three-dimensional volume image using the third plurality of corrected two-dimensional radiographic images; and

displaying, storing, or transmitting an image slice from the three-dimensional volume image.

2. The method of claim 1 wherein replacing the data for at least some of the segmented pixels comprises using image inpainting.

3. The method of claim 1 wherein segmenting pixels comprises automatically identifying seed pixels in one or more images obtained from the image stack.

4. The method of claim 1 wherein segmenting pixels comprises obtaining operator input for seed pixels for each of the metal objects and non-metal objects in one or more images obtained from the image stack.

5. The method of claim 1 wherein the first plurality of two-dimensional radiographic images are obtained from a cone-beam computed tomography system or obtained from a fan-beam computed tomography system, wherein the reconstructing the three-dimensional volume image comprises using filtered back projection.

6. The method of claim 1 wherein the combining further comprises subtracting data from the first plurality of two-dimensional image data.

7. The method of claim 1 , where segmenting pixels that represent metal objects from the image stack uses a repeated process of obtaining a two-dimensional image slice from the image stack, wherein the obtained two-dimensional image slice has image pixel data from a plurality of the first plurality of radiographic images, and processing the obtained two-dimensional image slice to identify metal features.

8. The method of claim 7 , where the replacing the data for at least some of the segmented pixels comprises using image inpainting.

9. A method for processing volume image data of a subject, the method executed at least in part on a computer, comprising:

obtaining a first plurality of two-dimensional radiographic images of the subject on an image detector, wherein each of the two-dimensional radiographic images is obtained with the image detector and a radiation source at a different scan angle relative to the subject;

arranging image data from the first plurality of two-dimensional radiographic images in an image stack so that corresponding pixel data from the detector is in register for each of the first plurality of two-dimensional radiographic images in the image stack;

segmenting pixels that represent metal objects from said each of the first plurality of the two-dimensional radiographic images of the image stack by a repeated process of obtaining an image stack slice from the image stack in a direction that is not parallel to the obtained two-dimensional radiographic images and processing the obtained image stack slice;

modifying data for at least some of the segmented pixels to generate a modified image stack with a reduced metal object content;

converting the modified image stack back into a second plurality of two-dimensional radiographic images at the different scan angles relative to the subject;

reconstructing a three-dimensional volume image according to the second plurality of two-dimensional radiographic images; and

displaying, storing, or transmitting an image slice from the reconstructed three-dimensional volume image.

10. The method of claim 9 , wherein modifying the data for at least some of the segmented pixels comprises using image inpainting, and wherein reconstructing the three-dimensional volume image comprises using filtered back projection.

11. The method of claim 9 , wherein segmenting pixels comprises automatically identifying seed pixels for each of the metal objects, bone objects and tissue objects in an image obtained from the image stack.

12. The method of claim 9 , wherein processing the obtained image stack slice comprises identifying pixels associated with metal features.

13. A method for processing volume image data of a subject, the method executed at least in part on a computer, comprising:

obtaining two-dimensional radiographic images of the subject on an image detector, where the two-dimensional radiographic images are obtained with the image detector and a radiation source at a different scan angle relative to the subject;

arranging image data from a first plurality of the two-dimensional radiographic images in an image stack so that corresponding pixel data from the detector is in register for each of the first plurality of two-dimensional radiographic images in the image stack;

re-arranging the image stack by a repeated process of obtaining an image stack slice from said each of the first plurality of the two-dimensional radiographic images of the image stack in a direction that is not parallel to the obtained two-dimensional radiographic images to form a second image stack;

replacing data for at least some pixels that represent metal objects in the second image stack by processing image stack slices in the second image stack;

obtaining a second plurality of modified two-dimensional radiographic images by obtaining a second image stack slice from the modified second image stack in a plane parallel to the obtained two-dimensional radiographic images;

combining some of the second plurality of modified two-dimensional radiographic images with the first plurality of two-dimensional radiographic images of the subject to form a third plurality of combined images;

reconstructing a three-dimensional volume image according to the third plurality of combined images; and

displaying, storing, or transmitting an image slice from the reconstructed three-dimensional volume image.

14. The method of claim 13 , where re-arranging the image stack executes a repeated process of obtaining the image stack slice from the image stack in a plane that is orthogonal to the plane of the obtained two-dimensional radiographic images in the image stack and processing the obtained image stack slice.