Methods and Systems for Registration of Images

• US 20080050043A1
• Filed: 08/20/2007
• Published: 02/28/2008
• Est. Priority Date: 08/22/2006
• Status: Active Grant

• Current Assignee: Siemens Healthineers AG (Siemens AG)
• Original Assignee: Siemens Medical Solutions USA Incorporated (Siemens AG)
• Inventors: Periaswamy, Senthil, Hermosillo Valadez, Gerardo
• Granted Patent: US 7,945,117 B2
• Time in Patent Office: Days
• US Class Current: 382/294
• CPC Class Codes: G06T 2207/10072 Tomographic images
• CPC Class Codes: G06T 2207/20221 Image fusion; Image merging
• CPC Class Codes: G06T 2207/30004 Biomedical image processing
• CPC Class Codes: G06T 2207/30096 Tumor; Lesion
• CPC Class Codes: G06T 7/0012 Biomedical image inspection
• CPC Class Codes: G06T 7/32 using correlation-based methods
• CPC Class Codes: G06T 7/38 Registration of image sequences
• CPC Class Codes: G06V 10/754 involving a deformation of the sample pattern or of the reference pattern; Elastic matching

Abstract

A method for registering images, for example medical images, includes acquiring a first image. A second image is acquired. A deformation field for registering the second image to the first image is calculated. The deformation field includes a sum of a plurality of weighted distortion vectors each of which has a substantially Gaussian distribution. The calculated deformation field is applied to the second image to register it to the first image.

Claims

1. A method for registering images, comprising:

acquiring a first image;

acquiring a second image;

calculating a deformation field for registering the second image to the first image, wherein the deformation field comprises a sum of weighted component deformation fields, each of which includes a distortion vector with an amplitude that is spatially distributed according to a substantially Gaussian distribution; and

applying the calculated deformation field to the second image to register it to the first image.

2. The method of claim 1, wherein the deformation field is defined by an equation comprising a set of variables that describe the plurality of distortion vectors, and calculating the deformation field comprises optimizing values for the set of variables to obtain the calculated deformation field.

3. The method of claim 2, wherein optimizing the values for the set of variables comprises adjusting the values to minimize an error metric representing a difference between an optimal deformation field and a deformation field calculated with a pre-optimized set of values.

4. The method of claim 2, wherein optimizing the values of the set of variables comprises computing the derivative of the error metric with respect to the set of variables.

5. The method of claim 3, wherein the error metric is selected according to the first or second images.

6. The method of claim 3, wherein the error metric is minimized using a conjugate gradient minimization technique.

7. The method of claim 3, wherein the error metric is minimized using a sum of squared differences technique.

13. The method of claim 11, wherein optimizing the values of the set of variables comprises computing the derivative of the error metric with respect to the set of variables.

14. The method of claim 13, wherein the error metric is selected according to the first or second images.

15. The method of claim 13, wherein the error metric is minimized using a conjugate gradient minimization technique.

16. The method of claim 13, wherein the error metric is minimized using a sum of squared differences technique.

8. A method for registering images within an image registration device, comprising:

acquiring a first image and a second image from a image database;

calculating a deformation field for registering the second image to the first image, wherein the deformation field comprises a sum of weighted component deformation fields, each of which includes a distortion vector with an amplitude that is spatially distributed according to a substantially Gaussian distribution; and

applying the calculated deformation field to the second image to register it to the first image; and

displaying the first image and the deformed second image on a display device.

9. A method for registering images, comprising:

acquiring a first image;

acquiring a second image;

calculating a distortion field for registering the second image to the first image, wherein the distortion field comprises a polynomial model; and

applying the calculated distortion field to the second image to register it to the first image.

10. The method of claim 9, wherein the distortion field comprising the polynomial model is a transformation matrix and applying the calculated distortion field to the second image comprises multiplying pixels of the second image by the transformation matrix to provide distorted pixels representing the pixel of the second image registered to the first image.

11. The method of claim 9, wherein the distortion field is defined by an equation comprising a set of variables that describe a transformation matrix, and calculating the distortion field comprises optimizing values for the set of variables to obtain the calculated distortion field.

12. The method of claim 11, wherein optimizing the values for the set of variables comprises adjusting the values to minimize an error metric representing a difference between an optimal distortion field and a distortion field calculated with a pre-optimized set of values.

17. A method for registering images within an image registration device, comprising:

acquiring a first image and a second image from a image database;

calculating a distortion field for registering the second image to the first image, wherein the distortion field comprises a polynomial model;

applying the calculated distortion field to the second image to register it to the first image; and

displaying the first image and the distorted second image on a display device.

18. A method for registering images, comprising:

acquiring a reference image;

acquiring a moving image;

calculating a gradient field for the reference image;

generating an incremental deformation based on the calculated gradient;

temporarily applying the generated incremental deformation to the reference image to determine whether the generated incremental deformation has increased a degree of registration between the reference and moving images; and

when the incremental deformation has increased the degree of registration between the images, applying an inverse of the incremental deformation to the moving image.

19. The method of claim 18, wherein the steps of generating an incremental deformation, temporarily applying the incremental deformation to the reference image and applying the inverse of the incremental deformation to the moving image when the degree of registration is increased, are repeated until the degree of registration is maximized.

20. The method of claim 19, wherein a similarity function is used to determine whether the degree of registration has been increased and to determine whether the degree of registration is maximized.

21. The method of claim 20, wherein the similarity function is a mean square error function representing a sum of absolute differences between the intensities of the moving image and the reference image.

22. The method of claim 18, wherein a gradient field for an image is calculated only once.

23. The method of claim 20, wherein the deformation field is initially unknown and is expressed according to the function:

φ

;

→

, the reference image I₁ is represented by a scalar function I₁;

→

and the moving image I₂ is represented by a scalar function I₂;

→

, and the similarity function is denoted by S, and wherein a displacement u({right arrow over (x)}) is solved for at each point x ∈

such that φ

({right arrow over (x)})={right arrow over (x)}+u({right arrow over (x)}).

24. The method of claim 20, wherein the similarity function is optimized uses the equation:

25. A method for registering images within an image registration device, comprising:

acquiring a first image and a second image from a image database;

acquiring a moving image;

calculating a gradient field for the reference image;

generating an incremental deformation based on the calculated gradient;

temporarily applying the generated incremental deformation to the reference image to determine whether the generated incremental deformation has increased a degree of registration between the reference and moving images;

when the incremental deformation has increased the degree of registration between the images, applying an inverse of the incremental deformation to the moving image; and

displaying the reference image and the moving image on a display device.