Image Processing

• US 20080166066A1
• Filed: 12/15/2005
• Published: 07/10/2008
• Est. Priority Date: 12/23/2004
• Status: Active Grant

• Current Assignee: Elekta AB
• Original Assignee: Elekta AB
• Inventors: Eriksson, Andréas Lars Gunnar
• Granted Patent: US 7,929,803 B2
• Time in Patent Office: Days
• US Class Current: 382/294
• CPC Class Codes: G06T 2207/30004 Biomedical image processing
• CPC Class Codes: G06T 7/30 Determination of transform parameters for the alignment of images, i.e. image registration
• CPC Class Codes: G06V 10/245 by locating a pattern; Special marks for positioning
• CPC Class Codes: G06V 2201/03 Recognition of patterns in medical or anatomical images

Abstract

Methods for the registration of images typically assume that there are only two images, an assumption that is not always valid. By using the remaining images to obtain a choice of paths between two selected images, the transformation between the two can be determined with greater accuracy by averaging those paths. When averaging the paths, greater weight can be given to paths whose accuracy is known or reasonably believed to be greatest. Iteration of the process may be possible, where the available computation power is available.

Claims

1. A method for determining a registration vector between two selected images of a group of at least three related images, the method comprising;

a. determining a set of registration vectors, each being between a respective pair of images in the group;

b. determining, from the set of registration vectors, at least one composite registration vector leading from one selected image to the other selected image via at least one image other than the selected images;

c. determining an optimized registration vector, being the average of a plurality of registration vectors selected from;

i. the registration vector between the selected images

ii. the at least one composite registration vector.

2. A method according to claim 1 including the step of estimating deviation errors in the plurality of registration vectors, and in which the average is weighted according to deviation error associated with each respective registration vector of the plurality.

3. A method according to claim 1 in which the one selected image is subsequently transformed according to the optimized registration vector.

4. A method according to claim 3 in which, subsequent to transformation, the one selected image is merged with the other selected image.

5. A software module for determining a registration vector between two selected images of a group of at least three related images, the module containing instructions to;

a. determine a set of registration vectors, each being between a respective pair of images in the group;

b. determine, from the set of registration vectors, at least one composite registration vector leading from one selected image to the other selected image via at least one image other than the selected images;

c. determine an optimized registration vector, being the average of a plurality of registration vectors selected from;

i. the registration vector between the selected images

ii. the at least one composite registration vector.

6. A module according to claim 5 including an instruction to estimate deviation errors in the plurality of registration vectors, and weight the average according to deviation error associated with each respective registration vector of the plurality.

7. A module according to claim 5 containing instructions to transform the one selected image according to the optimized registration vector.

8. A module according to claim 7 containing instructions to merge the one selected image with the other selected image, subsequent to transformation.

9. Image analysis apparatus arranged to determine a registration vector between two selected images of a group of at least three related images by a series of steps comprising;

a. determining a set of registration vectors, each being between a respective pair of images in the group;

b. determining, from the set of registration vectors, at least one composite registration vector leading from one selected image to the other selected image via at least one image other than the selected images;

c. determining an optimized registration vector, being the average of a plurality of registration vectors selected from;

i. the registration vector between the selected images

ii. the at least one composite registration vector.

10. Image analysis apparatus according to claim 9 further arranged to estimate deviation errors in the plurality of registration vectors, and to weight the average according to deviation error associated with each respective registration vector of the plurality.

11. Image analysis apparatus according to claim 9 arranged to transform the one selected image to the optimized registration vector.

12. Image analysis apparatus according to claim 11 arranged to merge the one selected image the other selected image, subsequent to transformation.

13. Image analysis apparatus according to claim 9 in which the images are images of a patient, derived from medical investigative apparatus.

14. (canceled)