Method and apparatus for semi-automatic segmentation technique for low-contrast tubular shaped objects

1. A method for segmenting anatomical objects from medical images, said method comprising using a computer to perform at least the following:

computing a centerline of an anatomical object;

deriving a set of reformatted images based on said centerline of said anatomical object;

receiving input from a user, said input being user drawn contours, said user drawn contours being drawn on a plurality of said reformatted images, said user drawn contours corresponding to the boundaries of the anatomical object to be segmented;

extracting said anatomical object based on said user drawn contours and displaying contours of the extracted volume; and,

editing the contours of the extracted volume.

2. The method of claim 1 , wherein the step of editing the contours of the extracted volume includes manual editing by a user.

3. The method of claim 2 , wherein the step of manually editing the contours of the extracted volume includes drawing the new contour, and detecting the portion of the current contour to be replaced by projecting end points of the new contour to the closest location on the current contour.

4. The method of claim 1 , wherein said reformatted images include a cross-section view orthogonal to the centerline and a longitudinal view tangent to the centerline.

5. The method of claim 1 , wherein said set of reformatted images includes an axial view, a sagittal view, or a coronal view.

6. The method of claim 1 , wherein said set of reformatted images includes a reformatted longitudinal view.

7. The method of claim 1 , wherein said set of reformatted images includes a curved reformatted view.

8. The method of claim 1 , wherein said set of reformatted images includes a lumen view.

9. The method of claim 1 , wherein said step of extracting said anatomical object based on said user drawn contours further includes utilizing previously computed segmentation results.

10. The method of claim 1 , wherein said step of extracting said anatomical object based on said user drawn contours further includes utilizing previous knowledge of the anatomy of a patient.

11. The method of claim 1 , wherein said anatomical object is a thrombus, said contours are displayed on cross-sectional reformatted slices, said reformatted slices are locally perpendicular to the centerline of the aorta lumen.

12. The method of claim 1 , wherein the step of editing the contours of the extracted volume includes pulling the current volume contour to the accurate location and applying a two-dimensional interpolation to said contour.

13. The method of claim 1 , wherein the step of editing the contours of the extracted volume includes applying a three-dimensional interpolation and computing a new volume for display.

14. A medical diagnostic imaging system, said system comprising:

a computer unit for manipulating data, said computer unit executing computer software for segmenting anatomical objects from medical images, said computer software computing a centerline of an anatomical object, deriving a set of reformatted images based on said centerline of said anatomical object, extracting said anatomical object based on said user drawn contours;

an input unit for receiving input from a user, said input being user drawn contours, said user drawn contours being drawn on a plurality of said reformatted images, said user drawn contours corresponding to the boundaries of the anatomical object to be segmented;

a display unit for displaying contours of the extracted volume.

15. The system of claim 14 , wherein said input unit receives input from a user, said input additionally being user input for manually editing the contours of the extracted volume.

16. The system of claim 14 , wherein said reformatted images include a cross-section view orthogonal to the centerline and a longitudinal view tangent to the centerline.

17. The system of claim 14 , wherein said set of reformatted images includes a curved reformatted view.

18. The system of claim 14 , wherein said set of reformatted images includes a reformatted longitudinal view.

19. The system of claim 14 , wherein said set of reformatted images includes a lumen view.

20. The system of claim 14 , wherein extracting said anatomical object based on said user drawn contours further includes utilizing previously computed segmentation results.

21. The system of claim 14 , wherein extracting said anatomical object based on said user drawn contours further includes utilizing previous knowledge of the anatomy of a patient.

22. The system of claim 14 , wherein said anatomical object is a thrombus, said contours are displayed on cross-sectional reformatted slices, said reformatted slices are locally perpendicular to the centerline of the aorta lumen.

23. A computer-readable storage medium including a set of instructions for a computer, the set of instructions comprising:

a computation routine for computing a centerline of an anatomical object;

a derivation routine for deriving a set of reformatted images based on said centerline of said object;

a receipt routine for receiving input from a user, said input being user drawn contours, said user drawn contours being drawn on a plurality of said reformatted images, said user drawn contours corresponding to the boundaries of the anatomical object to be segmented;

an extraction routine for extracting said anatomical object based on user drawn contours and displaying contours of the extracted volume; and

an editing routine for editing the contours of the extracted volume.