Spatially variant image deformation

• US 7,646,936 B2
• Filed: 10/03/2006
• Issued: 01/12/2010
• Est. Priority Date: 10/03/2006
• Status: Active Grant

• Current Assignee: Siemens Healthineers International AG
• Original Assignee: Varian Medical Systems International AG (Siemens AG)
• Inventors: Nord, Janne, Helminen, Hannu
• Primary Examiner(s): Kassa; Yosef
• Application Number:
  US11/542,952
• Publication Number:
  US 20080080788A1
• Time in Patent Office: 1,197 Days
• Field of Search: 382/128, 382/130, 382/131, 382/274, 382/275, 382/294
• US Class Current: 382/294
• CPC Class Codes: G06T 2207/30061 Lung
• CPC Class Codes: G06T 3/153 using elastic snapping
• CPC Class Codes: G06T 7/33 using feature-based methods

Abstract

Disclosed are systems for and methods of registering (i.e., aligning) a deformable image with a reference image subject to a spatially variant flexibility model and/or a non-Gaussian smoothing model. These systems and methods allow for the consideration of differences between the ways in which structures of interest may move within a patient. The registration includes modifying the deformable image to match similar features in the reference image. The systems include a deformation engine configured for performing a deformation algorithm subject to the spatially variant flexibility and/or smoothing models.

Claims

1. An image registration system comprising:

a deformation engine comprising;

first logic configured to compute a flexibility model comprising a first set of spatially variant constraints or spatially variant scalars and to apply the computed flexibility model to a medical deformable image to generate a set of displacement vectors, the set of displacement vectors being configured to apply a first force to the medical deformable image;

second logic configured to compute a non-Gaussian smoothing function and to apply the computed non-Gaussian smoothing function to the set of displacement vectors to create a smoothed set of displacement vectors;

third logic configured to compute a transform, the transform being configured to apply a second force to the medical deformable image, the second force opposing the first force; and

fourth logic configured to apply the smoothed set of displacement vectors and the transform to the medical deformable image to deform the medical deformable image to match a reference image.

2. The image registration system of claim 1, wherein the non-Gaussian smoothing function is spatially variant.

3. The image registration system of claim 1, further comprising an image generation apparatus configured to produce the medical deformable image.

4. The image registration system of claim 1, wherein the first set of spatially variant constraints or spatially variant scalars includes a first subset of constraints associated with a first direction of displacement between the reference image and the medical deformable image, and a second subset of constraints associated with a second direction of the displacement between the reference image and the medical deformable image.

5. The image registration system of claim 1, wherein the first set of spatially variant constraints or spatially variant scalars includes a plurality of constraints each associated with a different direction of displacement between the reference image and the medical deformable image.

6. The image registration system of claim 1, wherein the non-Gaussian smoothing function is an exponential function.

7. The image registration system of claim 1, wherein the exponential function includes a set of spatially variant decay parameters.

8. The image registration system of claim 1, wherein the non-Gaussian smoothing function is configured to be applied using a multidimensional matrix.

9. An image registration system comprising:

a deformation engine comprising;

first logic configured to compute a flexibility model comprising a set of spatially variant constraints or spatially variant scalars and to apply the computed flexibility model to a medical deformable image to generate a set of displacement vectors, the set of displacement vectors being configured to apply a first force to the medical deformable image;

second logic configured to compute a Gaussian smoothing function and to apply the computed Gaussian smoothing function to the set of displacement vectors to create a smoothed set of displacement vectors;

third logic configured to compute a transform, the transform being configured to apply a second force to the medical deformable image, the second force opposing the first force; and

fourth logic configured to apply the smoothed set of displacement vectors and the transform to the medical deformable image to deform the medical deformable image to match a reference image.

10. The image registration system of claim 9, wherein the set of spatially variant constraints includes a first subset of constraints associated with one direction and a second subset of constraints associated with a second direction.

11. An image registration system comprising:

a deformation engine comprising;

first logic configured to compute a flexibility model and to apply the computed flexibility model to a medical deformable image to generate a set of displacement vectors, the set of displacement vectors being configured to apply a first force to the medical deformable image;

second logic configured to compute a spatially variant smoothing function and to apply the computed spatially variant smoothing function to the set of displacement vectors to create a smoothed set of displacement vectors;

third logic configured to compute a transform, the transform being configured to apply a second force to the medical deformable image, the second force opposing the first force; and

fourth logic configured to apply the set of displacement vectors and the transform to the medical deformable image to deforming the medical deformable image to match a reference image.

12. The image registration system of claim 11, wherein the spatially variant smoothing function includes an exponential function.

13. The image registration system of claim 12, wherein the exponential function is dependent on a set of directionally dependent decay parameters.

14. The image registration system of claim 12, wherein the spatially variant smoothing function is configured to be applied using a multi-dimensional matrix.

15. A method comprising:

computing a set of displacement vectors according to a flexibility model comprising a first set of spatially variant constraints or spatially variant scalars, the set of displacement vectors being configured to apply a first force to a medical deformable image;

applying a non-Gaussian smoothing function to the set of displacement vectors to create a smoothed set of displacement vectors;

computing a transform configured to apply a second force to the medical deformable image, the second force opposing the first force; and

applying the transform and the smoothed set of displacement vectors to the medical deformable image to match the medical deformable image to a reference image.

16. The method of claim 15, wherein the non-Gaussian smoothing function is spatially variant.

17. The method of claim 15, wherein the first set of spatially variant constraints or spatially variant scalars includes a first subset of constraints or scalars associated with a first direction of displacement between the reference image and the medical deformable image, and a second subset of constraints or scalars associated with a second direction of the displacement between the reference image and the medical deformable image.

18. The method of claim 15, wherein the first set of spatially variant constraints or spatially variant scalars includes a plurality of constraints or scalars each associated with a different direction of displacement between the reference image and the medical deformable image.

19. The method of claim 15, wherein the non-Gaussian smoothing function includes an exponential function.

20. The method of claim 19, wherein the exponential function includes a set of spatially variant decay parameters.

21. The method of claim 15, wherein the computing of the non-Gaussian smoothing function includes using a multidimensional matrix.

22. The method of claim 15, wherein the first set of spatially variant constraints or spatially variant scalars includes a plurality of spatially variant multi-dimensional constraints.

23. A method comprising:

computing a set of displacement vectors according to a flexibility model comprising a set of spatially variant constraints or spatially variant scalars, the set of displacement vectors being configured to apply a first force to a medical deformable image;

applying a Gaussian smoothing function to the set of displacement vectors to create a smoothed set of displacement vectors;

computing a transform configured to apply a second force to the medical deformable image, the second force opposing the first force; and

applying the computed transform and the smoothed set of displacement vectors to the medical deformable image to match the medical deformable image to a reference image.

24. The method of claim 23, wherein the set of spatially variant constraints includes a first subset of constraints associated with one direction and a second subset of constraints associated with a second direction.

25. A method comprising:

computing a set of displacement vectors according to a flexibility model and applying the set of displacement vectors to a medical deformable image, the set of displacement vectors being configured to apply a first force to the medical deformable image;

applying a spatially variant smoothing function to the medical deformable image;

computing a transform configured to apply a second force to the medical deformable image, the second force opposing the first force; and

applying the transform to the medical deformable image to match the medical deformable image to a reference image.

26. The method of claim 25, wherein the spatially variant smoothing function includes an exponential function.

27. The method of claim 25, wherein the spatially variant smoothing function is applied using a multi-dimensional matrix.