Device and method of computing a transformation linking two images

• US 20020054699A1
• Filed: 09/25/2001
• Published: 05/09/2002
• Est. Priority Date: 09/26/2000
• Status: Active Grant

• Current Assignee: Koninklijke Philips Electronics N.V. (Koninklijke Philips N.V.)
• Original Assignee: Koninklijke Philips Electronics N.V. (Koninklijke Philips N.V.)
• Inventors: Weese, Juergen, Roesch, Peter
• Granted Patent: US 6,950,542 B2
• Time in Patent Office: Days
• US Class Current: 382/131
• CPC Class Codes: G06V 10/754 involving a deformation of the sample pattern or of the reference pattern; Elastic matching

Abstract

The invention relates to a method of computing the transformation which transforms two different images (10, 10′) of an object one into the other while describing the motion or deformation of the object. In accordance with the method first the local transformation parameters are computed for sub-regions (1, 2, 3, 4, 5). By choosing the subregions so as to be sufficiently small, a rigid transformation can be used for this purpose. Starting from at least one first predetermined sub-region (1), the local transformation parameters of the neighboring sub-regions (2, 3, 4, 5) are successively computed, the starting values on which each computation is based being the already determined local transformation parameters of neighboring sub-regions. Using the local transformation parameters thus determined for sub-regions, the overall transformation can subsequently be computed.

Claims

1. A method of computing the transformation which transforms two images (10, 10′

) of an object, preferably acquired by means of computed tomography or magnetic resonance tomography, one into the other, first the local transformation parameters being successively computed for sub-regions (1, 2, 3, 4, 5) of the images and subsequently the overall transformation being computed therefrom, characterized in that the successive computation of the local transformation parameters for the sub-regions is performed as follows;

a) the local transformation parameters are predetermined or computed for at least one first sub-region (1);

b) at least one sub-region (2) that has not yet been computed is formed or selected from among a quantity of predetermined sub-regions (2, 3, 4, 5);

c) for each sub-region (2) in the step b) the local transformation parameters are calculated, the local transformation parameters of at least one neighboring sub-region (1) that has already been successfully treated being used as starting values;

d) the steps b) and c) are repeated a number of times until the local transformation parameters have been computed for an adequate distribution of sub-regions.

2. A method as claimed in claim 1, characterized in that the sub-regions (1, 2, 3, 4, 5) are chosen so as to be sufficiently small so that the local transformation parameters may be assumed to be constant for all points of the sub-region while maintaining a predetermined accuracy.

3. A method as claimed in claim 1 or 2, characterized in that exclusively subregions (1, 2, 3, 4, 5) that contain adequately strong structuring are used.

4. A method as claimed in at least one of the claims 1 to 3, characterized in that the sub-regions are chosen so as to overlap at least partly.

5. A method as claimed in at least one of the claims 1 to 4, characterized in that prior to the step a) a suitable distribution of sub-regions (1, 2, 3, 4, 5) is determined in one of the images (10), and that in the step b) at least one sub-region (2) that has not yet been computed is selected from said distribution.

6. A method as claimed in claim 5, characterized in that the sub-region (2) that is selected in the step b) is the sub-region that has the highest weighting factor from among all sub-regions (2, 3, 4, 5) that have not yet been computed, a high weight being induced by a small distance from an already registered sub-region and/or a high accuracy of the transformation parameters determined for the already registered sub-region.

7. A method as claimed in at least one of the claims 1 to 6, characterized in that the local transformation parameters computed for a sub-region (2) are used further exclusively if their accuracy exceeds a specified threshold value.

8. A method as claimed in at least one of the claims 1 to 7, characterized in that in the step c) the starting value of the local transformation parameters is determined as a mean value of the already computed local transformation parameters that is weighted with the distance of the associated sub-regions (1) and/or with their registration accuracy.

9. A computer program for computing the transformation which transforms two digitized images (10, 10′

) of an object, preferably acquired by means of computed tomography or magnetic resonance tomography, one into the other, characterized in that the computation is performed by means of a method as claimed in one of the claims 1 to 8.

10. A device for computing the transformation which transforms two digitized images (10, 10′

) of an object, preferably acquired by means of computed tomography or magnetic resonance tomography, one into the other, comprising a central processing unit that is connected to at least one memory unit for the images and for a computer program executable by the central processing unit, the program being adapted to control the calculation of the transformation according to a method as claimed in one of the claims 1 to 9.