System and Method for Segmenting a Region in a Medical Image

• US 20080267468A1
• Filed: 10/10/2007
• Published: 10/30/2008
• Est. Priority Date: 10/10/2006
• Status: Active Grant

• Current Assignee: Merative US LP
• Original Assignee: Merge Healthcare Canada Corporation
• Inventors: Accomazzi, Vittorio, GEIGER, Paul
• Granted Patent: US 8,036,436 B2
• Time in Patent Office: Days
• US Class Current: 382/128
• CPC Class Codes: A61B 8/13 Tomography A61B8/10, A61B8/12 take precedence; tomography for radiation diagnosis A61B6/02
• CPC Class Codes: A61B 8/469 for selection of a region of interest
• CPC Class Codes: G01S 15/8993 Three dimensional imaging systems
• CPC Class Codes: G01S 7/52036 using analysis of echo signal for target characterisation
• CPC Class Codes: G01S 7/52074 Composite displays, e.g. split-screen displays; Combination of multiple images or of images and alphanumeric tabular information
• CPC Class Codes: G01S 7/52084 related to particular user interfaces special user input means for ultrasonic medical diagnostic devices A61B8/467
• CPC Class Codes: G06T 2207/10132 Ultrasound image
• CPC Class Codes: G06T 2207/20092 Interactive image processing based on input by user
• CPC Class Codes: G06T 7/12 Edge-based segmentation
• CPC Class Codes: G06T 7/149 involving deformable models, e.g. active contour models
• CPC Class Codes: G06V 10/248 by interactive preprocessing or interactive shape modelling, e.g. feature points assigned by a user
• CPC Class Codes: G06V 10/457 by analysing connectivity, e.g. edge linking, connected component analysis or slices
• CPC Class Codes: G06V 2201/031 of internal organs

Abstract

A method and computer program are provided for segmenting a prostate from a medical image such as an ultrasound image. Each slice in a study is analysed and uses a manual initialization to obtain an initial contour, in which a user selects initial points that are automatically rendered on the image. An automatic refinement stage then snaps the contour to the prostate boundary in the image based on a pre-stored anatomical atlas and edge information obtained from the image. A manual adjustment stage may then be performed and if selected, the contour is automatically propagated to the next image slice to avoid the manual initialization. An auxiliary image slice may be referred to, which indicates how the prostate shape changes from slice to slices e.g. by providing a perpendicular profile.

Claims

1. A method for segmenting a region in an image comprising generating an initial contour from a plurality of initial points by interpolating a contour segment between each pair of neighbouring ones of said plurality of initial points using one or more control points placed relative to said each pair;

and adjusting the positions of said control points to modify each said contour segment to refine said initial contour according to information regarding an expected shape for said region.

2. The method according to claim 1 wherein said interpolating comprises using a spline-based interpolation between said each pair wherein said control points are located outside of said initial contour.

3. The method according to claim 1 comprising examining boundary information in said image to generate a score, refining said initial contour a plurality of times to produce a plurality of scores and generating a refined contour according to the highest of said scores.

4. The method according to claim 3 wherein said examining boundary information comprises comparing said initial contour to boundary information in said image to generate a first score;

comparing said initial contour to said information to generate a second score, said information being indicative of an expected shape for said region;

combining said first and second scores to obtain a combined score;

reshaping said initial contour a plurality of times to obtain a plurality of adjusted contours;

generating said combined score for each said plurality of adjusted contours; and

generating said refined contour according to the highest of said combined scores.

5. The method according to claim 4 wherein said first score is calculated by incrementing by the value of each pixel in a gradient-magnitude image such that said first score is high when said contour is close to edges in said image.

6. The method according to claim 4 wherein said second score is calculated according to how close said contour is from an average shape, said average shape being determined from a plurality of shapes in an empirically derived atlas of expected shapes.

7. The method according to claim 4 wherein said combined score is computed as a weighted combination of said first and second scores.

8. The method according to claim 1 comprising generating said plurality of initial points by enabling the selection of a first point at a first extent of said region;

enabling the selection of a second point at a second extent of said region being opposite said first point to define a line of symmetry;

generating a provisional contour passing through said first and second points;

providing a third point on said provisional contour on one side of said line of symmetry;

enabling movement of said third point towards a third extent of said region; and

during movement of said third point, automatically moving a fourth point on said provisional contour on the other side of said line of symmetry in a direction opposite that of said third point towards a fourth extent of said region.

9. The method according to claim 1 comprising obtaining at least one auxiliary image slice of said region at a different angle than a stack of image slices containing said region to provide a profile of said region in another dimension;

repeating said method for each said image slices and referencing said profile to approximate a next segmentation contour for a next one of said image slices while propagating through said stack of image slices.

10. The method according to claim 1 wherein said information is provided by any one or more of the following:

functional knowledge derived from medical images, geometry of said region, symmetry of said region, real-time information acquired during imaging, an empirically derived atlas of expected shapes, and localization information.

11. The method according to claim 1 wherein said control points are modified a plurality of times and at each modification, a plausibility of the shape of said region is determined from an atlas of empirically derived expected shapes for said region, said plausibility being used to choose a refined contour.

12. A computer readable medium comprising computer executable instructions for causing an image processing device to segment a region in an image by generating an initial contour from a plurality of initial points by interpolating a contour segment between each pair of neighbouring ones of said plurality of initial points using one or more control points placed relative to said each pair;

and adjusting the positions of said control points to modify each said contour segment to refine said initial contour according to known information regarding an expected shape for said region.

13. The computer readable medium according to claim 12 wherein said interpolating comprises using a spline-based interpolation between said each pair wherein said control points are located outside of said initial contour.

14. The computer readable medium according to claim 12 comprising instructions for examining boundary information in said image to generate a score, refining said initial contour a plurality of times to produce a plurality of scores and generating a refined contour according to the highest of said scores.

15. The computer readable medium according to claim 14 wherein said examining boundary information comprises comparing said initial contour to boundary information in said image to generate a first score;

comparing said initial contour to said information to generate a second score, said information being indicative of an expected shape for said region;

combining said first and second scores to obtain a combined score;

reshaping said initial contour a plurality of times to obtain a plurality of adjusted contours;

generating said combined score for each said plurality of adjusted contours; and

generating said refined contour according to the highest of said combined scores.

16. The computer readable medium according to claim 15 wherein said first score is calculated by incrementing by the value of each pixel in a gradient-magnitude image such that said first score is high when said contour is close to edges in said image.

17. The computer readable medium according to claim 15 wherein said second score is calculated according to how close said contour is from an average shape, said average shape being determined from a plurality of shapes in an empirically derived atlas of expected shapes.

18. The computer readable medium according to claim 15 wherein said combined score is computed as a weighted combination of said first and second scores.

19. The computer readable medium according to claim 12 comprising instructions for generating said plurality of initial points by enabling the selection of a first point at a first extent of said region;

enabling the selection of a second point at a second extent of said region being opposite said first point to define a line of symmetry;

generating a provisional contour passing through said first and second points;

providing a third point on said provisional contour on one side of said line of symmetry;

enabling movement of said third point towards a third extent of said region; and

during movement of said third point, automatically moving a fourth point on said provisional contour on the other side of said line of symmetry in a direction opposite that of said third point towards a fourth extent of said region.

20. The computer readable medium according to claim 12 comprising instructions for obtaining at least one auxiliary image slice of said region at a different angle than a stack of image slices containing said region to provide a profile of said region in another dimension;

repeating said method for each said image slices and referencing said profile to approximate a next segmentation contour for a next one of said image slices while propagating through said stack of image slices.

21. The computer readable medium according to claim 12 wherein said information is provided by any one or more of the following:

functional knowledge derived from medical images, geometry of said region, symmetry of said region, real-time information acquired during imaging, an empirically derived atlas of expected shapes, and localization information.

22. The computer readable medium according to claim 12 wherein said control points are modified a plurality of times and at each modification, a plausibility of the shape of said region is determined from an atlas of empirically derived expected shapes for said region, said plausibility being used to choose a refined contour.

23. A method for generating, a initial contour for segmenting a region in an image comprising enabling the selection of a first point at a first extent of said region;

enabling the selection of a second point at a second extent of said region being opposite said first point to define a line of symmetry;

generating a provisional contour passing through said first and second points;

providing a third point on said provisional contour on one side of said line of symmetry;

enabling movement of said third point towards a third extent of said region;

during movement of said third point, automatically moving a fourth point on said provisional contour on the other side of said line of symmetry in a direction opposite that of said third point towards a fourth extent of said region; and

interpolating between said points according to information indicative of an expected shape for said region to generate said initial contour.

24. A method for refining a segmentation contour generated on an image for a region in said image comprising comparing said contour to boundary information in said image to generate a first score;

comparing said contour to information indicative of an expected shape for said region to generate a second score;

combining said first and second scores to obtain a combined score;

reshaping said contour a plurality of times to obtain a plurality of adjusted contours;

generating said combined score for each said plurality of adjusted contours; and

generating a refined contour according to the highest of said combined scores.

25. A method for segmenting a three dimensional region from a stack of image slices of said region comprising obtaining at least one auxiliary image slice of said region at a different angle than said stack to provide a profile of said region in another dimension;

and referencing said profile to approximate a next segmentation contour for a next one of said image slices while propagating through said stack of image slices.