Layer reconstruction from dual-energy image pairs

• US 8,005,288 B2
• Filed: 04/11/2008
• Issued: 08/23/2011
• Est. Priority Date: 04/24/2007
• Status: Expired due to Fees

• Current Assignee: Siemens Healthcare GMBH (Siemens AG)
• Original Assignee: Siemens AG
• Inventors: Fang, Tong, Chen, Yunqiang
• Primary Examiner(s): Werner; Brian
• Assistant Examiner(s): SHAH, UTPAL D
• Application Number:
  US12/101,487
• Publication Number:
  US 20080267474A1
• Time in Patent Office: 1,229 Days
• Field of Search: 382/132, 382/284, 382/294
• US Class Current: 382/132
• CPC Class Codes: G06T 2207/10116 X-ray image
• CPC Class Codes: G06T 2207/10144 Varying exposure
• CPC Class Codes: G06T 2207/30004 Biomedical image processing
• CPC Class Codes: G06T 5/50 using two or more images, e.g. averaging or subtraction
• CPC Class Codes: G06T 5/77 Retouching; Inpainting; Scratch removal
• CPC Class Codes: G06T 7/33 using feature-based methods

Abstract

A system and method for layer reconstruction from dual-energy image pairs are provided, the method including: receiving a pair of dual-energy images, one having a relatively high energy dose and the other having a relatively low energy dose; ascertaining that a first relatively motionless layer is substantially aligned between the high and low dose images; computing a preliminary image of a second layer that has non-rigid motion relative to the first layer; detecting the relative motion of the second layer relative to the first layer; generating a mask in accordance with the detected motion; filling the motion area corresponding to the mask with gradients of the high-dose image; removing the first layer; and inpainting the motion area.

Claims

1. A method for layer reconstruction from dual-energy image pairs, the method comprising:

receiving a pair of images including a first image captured at a relatively high-energy dose and a second image captured at a relatively low energy dose;

determining gradients aligned between the pair of images as a first layer;

determining a second layer by subtracting the first layer from a gradient of the first image;

detecting a motion area of the second layer relative to the first layer;

inpainting the motion area of the second layer with gradients of the first image; and

reconstructing the second layer by removing the gradients of the first layer from the gradients of the first image inpainted in the motion area of the second layer, wherein the second layer is an image of soft tissue.

2. The method of claim 1 wherein detecting the motion area of the second layer relative to the first layer comprises determining a normalized correlation between the gradient of the first image and a gradient of the second image, respectively.

3. The method of claim 1 further comprising generating a mask in accordance with the motion area, wherein generating the mask comprises:

determining regions of relative correlation between the pair of images;

determining areas where a gradient of the second layer is high; and

activating at least one pixel of the mask that is both close to a low-correlation region and near a high-gradient region.

4. The method of claim 1 further comprising generating a mask in accordance with the motion area, wherein generating the mask comprises:

correlating the gradient of the first image with gradient of the second image;

determining a gradient of the second layer corresponding to a non-correlation between a scaled version of the gradient of the first image and the gradient of the second image;

determining the mask in accordance with the correlation and the gradient of the second layer.

5. The method of claim 1 further comprising reconstructing the first layer by removing the gradients of a reconstructed second layer from the gradients of the first image, wherein the second layer is an image of bone.

6. A non-transitory program storage device readable by computer, tangibly embodying a program of instructions executable by the computer to perform program steps for layer reconstruction from dual-energy image pairs, the program steps comprising:

receiving a pair of images including a first image captured at a relatively high-energy dose and a second image captured at a relatively low energy dose;

determining gradients aligned between the pair of images as a first layer;

determining a second layer by subtracting the first layer from a gradient of the first image;

detecting a motion area of the second layer relative to the first layer;

inpainting the motion area of the second layer with gradients of the first image; and

reconstructing the second layer by removing the gradients of the first layer from the gradients of the first image inpainted in the motion area of the second layer.

7. The device of claim 6 wherein detecting the motion area of the second layer relative to the first layer comprises determining a normalized correlation between the gradient of the first image and a gradient of the second image, respectively.

8. The device of claim 6 further comprising generating a mask in accordance with the motion area, wherein generating the mask comprises:

determining regions of relative correlation between the pair of images;

determining areas where a gradient of the second layer is high; and

activating at least one pixel of the mask that is both close to a low-correlation region and near a high-gradient region.

9. The device of claim 6 further comprising generating a mask in accordance with the motion area, wherein generating the mask comprises:

correlating the gradient of the first image with a gradient of the second image;

determining a gradient of the second layer corresponding to a non-correlation between a scaled version of the gradient of the first image and the gradient of the second image;

determining the mask in accordance with the correlation and the gradient of the second layer.

10. The method of claim 6 further comprising reconstructing the first layer by removing the gradients of a reconstructed second layer from the gradients of the first image.