Method and apparatus for image registration using large deformation diffeomorphisms on a sphere

• US 6,633,686 B1
• Filed: 09/20/2000
• Issued: 10/14/2003
• Est. Priority Date: 11/05/1998
• Status: Expired due to Term

• Current Assignee: Washington University In St Louis
• Original Assignee: Washington University In St Louis
• Inventors: Joshi, Sarang, Bakircioglu, Muge M., Miller, Michael I.
• Primary Examiner(s): Mariam, Daniel G.
• Application Number:
  US09/665,688
• Time in Patent Office: 1,119 Days
• Field of Search: 382/131, 382/151, 382/201, 382/209, 382/215, 382/217, 382/218, 382/276, 382/280, 382/282, 382/291, 382/294, 345/424, 345/427, 345/418, 378/4, 378/13, 378/20, 600/424, 600/425, 600/426
• US Class Current: 382/294
• CPC Class Codes: G06T 2207/30004 Biomedical image processing
• CPC Class Codes: G06T 3/153 using elastic snapping
• CPC Class Codes: G06T 7/33 using feature-based methods
• CPC Class Codes: G06T 7/37 using transform domain methods
• CPC Class Codes: G06T 7/70 Determining position or orientation of objects or cameras camera calibration G06T7/80
• CPC Class Codes: G06V 10/754 involving a deformation of the sample pattern or of the reference pattern; Elastic matching

Abstract

An apparatus and method for image registration of a template image with a target image with large deformation. The apparatus and method involve computing a large deformation transform based on landmark manifolds, image data or both. The apparatus and method are capable of registering images with a small number of landmark points. Registering the images is accomplished by applying the large deformation transform.

Claims

1. A method for registering a template image with a target image, the template image containing a plurality of points and the target image containing a plurality of points, comprising:

defining at least one manifold landmark point in said template image;

identifying at least one point in said target image corresponding to said defined at least one manifold landmark point;

selecting a coordinate frame suitable for spherical geometries;

computing a large deformation transform using said selected coordinate frame, at least one manifold landmark transformation operator, and at least one manifold landmark transformation boundary value, said large deformation transform relating said at least one manifold landmark point in said template image to said corresponding at least one point in said target image; and

registering said template image with said target image using said large deformation transform in the selected coordinate frame.

2. The method of claim 1, wherein the step of computing a large deformation transform further includes the substep of computing a diffeomorphic transform.

3. The method of claim 1, wherein the step of computing a large deformation transform further includes the substep of accessing at least one precomputed transform value.

4. The method of claim 1, wherein the step of computing a large deformation transform further includes the substep of using a fast Fourier transform.

5. The method of claim 1, wherein the step of defining at least one manifold landmark point in said template image includes the substep of:

performing at least one of;

defining at least one manifold landmark point of dimension greater than zero in said template image;

defining at least one point in said template image;

defining points of a curve in said template image;

defining points of a surface in said template image;

or defining points of a volume in said template image.

6. The method of claim 1, wherein the step of computing a large deformation transform includes the substep of:

using at least one of a linear differentiable operator, a periodic boundary value, an infinite boundary value, mixed Dirichlet and Neumann boundary values, a Neumann boundary value, or a Dirichlet boundary value.

7. A method for registering a template image with a target image, comprising:

selecting a coordinate frame which is two-dimensional only and which is suitable for spherical geometries; and

registering the target and template image using a large deformation transform in the selected coordinate frame.

8. The method of claim 7, wherein the step of registering the target and template image further includes the substep of computing a diffeomorphic transform.

9. The method of claim 7, wherein the step of registering the target and template image further includes the substep of accessing at least one precomputed transform value.

10. The method of claim 7, wherein the step of registering the target and template image further includes the substep of using a fast Fourier transform.

11. The method of claim 7 further including, prior to registering the target and template image:

performing at least one of;

defining at least one manifold landmark point of dimension greater than zero in said template image;

defining at least one point in said template image;

defining points of a curve in said template image;

defining points of a surface in said template image;

or defining points of a volume in said template image.

12. A method for registering a template image with a target image, wherein the template image contains a plurality of points and the target image contains a plurality of points, comprising:

selecting a coordinate frame suitable for spherical geometries;

computing a first large deformation transform using at least one manifold landmark transformation operator relating at least one manifold landmark point in said template image to at least one corresponding point in said target image;

computing a second large deformation transform using at least one image transformation operator relating said template image to said target image;

fusing said first and second large deformation transforms; and

registering said template image with said target image using said fused large deformation transforms in said selected coordinate frame.

13. An apparatus for registering a template image with a target image, the template image containing a plurality of points and the target image containing a plurality of points, comprising:

means for defining at least one manifold landmark point in said template image;

means for identifying at least one point in said target image corresponding to said defined at least one manifold landmark point;

means for selecting a coordinate frame suitable for spherical geometries;

means for computing a large deformation transform using said selected coordinate frame, at least one manifold landmark transformation operator, and at least one manifold landmark transformation boundary value, said large deformation transform relating said at least one manifold landmark point in said template image to said corresponding at least one point in said target image; and

means for registering said template image with said target image using said large deformation transform in the selected coordinate frame.

14. An apparatus for registering a template image with a target image, the template image containing a plurality of points and the target image containing a plurality of points, comprising:

means for selecting a coordinate frame which is two-dimensional only and which is suitable for spherical geometries; and

means for registering the target and template image using a large deformation transform in the selected coordinate frame.

15. An apparatus for registering a template image with a target image, the template image containing a plurality of points and the target image containing a plurality of points, comprising:

a coordinate frame which is two-dimensional only and which is suitable for spherical geometries; and

a registration processor for registering the target and template image using a large deformation transform in the selected coordinate frame.

16. An apparatus for registering a template image with a target image, wherein the template image contains a plurality of points and the target image contains a plurality of points, comprising:

a pointing device for defining at least one manifold landmark point in said template image and for identifying at least one point in said target image corresponding to said at least one defined manifold landmark point;

a coordinate frame selector for selecting a coordinate frame suitable for spherical geometries;

a first data processing unit for computing a first large deformation transform using a manifold landmark transformation operator relating said at least one manifold landmark point in said template image to said corresponding at least one point in said target image;

a distance quantifier for defining a distance between said target image and said template image;

a second data processing unit for computing a second large deformation transform using an image transformation operator relating said template image to said target image; and

a third data processing unit for fusing said first and second large deformation transforms; and

a fourth data processing unit for registering said template image with said target image using said fused large deformation transforms in said coordinate frame suitable for spherical geometries.

17. A computer program product for use in a computer adapted for registering a template image with a target image, the computer program product comprising a computer readable medium for storing computer readable code means, which when executed by the computer, enables the computer to register a template image with a target image, and wherein the computer readable code means includes computer readable instructions for causing the computer to execute a method comprising:

selecting a coordinate frame which is two-dimensional only and which is suitable for spherical geometries; and

registering the target and template image using a large deformation transform in the selected coordinate frame.