System and method for tracking a global shape of an object in motion

• US 7,558,402 B2
• Filed: 03/05/2004
• Issued: 07/07/2009
• Est. Priority Date: 03/07/2003
• Status: Expired due to Fees

• Current Assignee: Siemens Medical Solutions USA Incorporated (Siemens AG)
• Original Assignee: Siemens Medical Solutions USA Incorporated (Siemens AG)
• Inventors: Krishnan, Sriram, Georgescu, Bogdan, Zhou, Xiang Sean, Comaniciu, Dorin
• Primary Examiner(s): Mehta; Bhavesh M
• Assistant Examiner(s): ZARKA, DAVID PETER
• Application Number:
  US10/794,476
• Publication Number:
  US 20040208341A1
• Time in Patent Office: 1,950 Days
• Field of Search: 382/103
• US Class Current: 382/103
• CPC Class Codes: G06F 18/2135 based on approximation criteria, e.g. principal component analysis
• CPC Class Codes: G06F 18/253 of extracted features
• CPC Class Codes: G06T 2207/10132 Ultrasound image
• CPC Class Codes: G06T 2207/30048 Heart; Cardiac
• CPC Class Codes: G06T 7/215 Motion-based segmentation
• CPC Class Codes: G06T 7/251 involving models
• CPC Class Codes: G06V 10/245 by locating a pattern; Special marks for positioning
• CPC Class Codes: G06V 10/7715 Feature extraction, e.g. by transforming the feature space, e.g. multi-dimensional scaling [MDS]; Mappings, e.g. subspace methods
• CPC Class Codes: G06V 10/806 of extracted features
• CPC Class Codes: G06V 2201/031 of internal organs

Abstract

A system and method for tracking a global shape of an object in motion is disclosed. One or more control points along an initial contour of the global shape are defined. Each of the one or more control points is tracked as the object is in motion. Uncertainty of a location of a control point in motion is represented using a number of techniques. The uncertainty to constrain the global shape is exploited using a prior shape model. In an alternative embodiment, multiple appearance models are built for each control point and the motion vectors produced by each model are combined in order to track the shape of the object. The movement of the shape of the object can be visually tracked using a display and color vectors.

Claims

1. A computer implemented method for tracking a deformable global shape of an object in motion in a sequence of digital medical images, the method performed by the computer comprising the steps of:

providing a probabilistic subspace shape model for the deformable global shape derived from a set of training shapes;

defining one or more control points along the deformable global shape;

tracking each of the one or more control points as the object is in motion;

estimating uncertainty of a location of a control point in motion;

exploiting the uncertainty to constrain the deformable global shape using a point dependent covariance matrix to employ the subspace shape constraint model using a non-orthogonal projection;

fusing the subspace constraint model and the uncertainty to determine a current location of the object by correcting outlier control points according to the probabilistic subspace shape model; and

remedying error accumulation during tracking of the one or more control points by using a multi-template adaptive matching framework.

2. The method of claim 1 wherein the step of defining the one or more control points is performed manually.

3. The method of claim 1 wherein the step of defining the one or more control points is performed automatically.

4. The method of claim 1 wherein the step of tracking each of the one or more control points employs a Bayesian kernel matching approach.

5. The method of claim 1 wherein the step of tracking each of the one or more control points employs an optical flow based approach.

7. The method of claim 1 wherein the subspace shape constraint model is an eigenspace model.

8. The method of claim 1 wherein the object is a myocardial wall of a left ventricle.

9. The method of claim 8 wherein the myocardial wall is imaged via an echocardiogram.

10. The method of claim 8 wherein the shape tracking is used to detect abnormalities in the myocardial wall.

11. The method of claim 8 wherein the myocardial wall is an endocardial wall.

12. The method of claim 8 wherein the myocardial wall is an epicardial wall.

13. The method of claim 1 wherein each control point is tracked independently of the other control points.

14. The method of claim 1 wherein the step of representing uncertainty includes measuring anisotropic uncertainty.

15. The method of claim 1 wherein a strongly adapted Principal Component Analysis (SA-PCA) model is used to constrain the shape.

16. The method of claim 15 wherein the initial shape is used to construct the SA-PCA model.

17. The method of claim 1 wherein the global shape is defined by an ordered set of control points.

18. The method of claim 1 wherein the prior shape model is derived from training data.

19. The method of claim 1 wherein the shape of the object in motion is deformable.

20. The method of claim 19 wherein the motion is non-rigid.

21. The method of claim 1 further comprising the step of:

displaying each subsequent shape.

6. The method of 1 wherein the subspace shape constraint model is a Gaussian model.

22. A program storage device readable by a computer, tangibly embodying a program of instructions executable by the computer to perform the method steps for tracking a global shape of an object in motion in a sequence of digital medical images, method comprising:

providing a probabilistic subspace shape model for the deformable global shape derived from a set training shapes;

defining one or more control points along an initial contour of the deformable global shape;

tracking each of the one or more control points as the object is in motion;

estimating uncertainty of a location of a control point in motion;

exploiting the uncertainty to constrain the deformable global shape using a point dependent covariance matrix to employ the subspace shape constraint model using a non-orthogonal projection;

fusing the subspace constraint model and the uncertainty to determine a current location of the object by correcting outlier control points according to the probabilistic subspace shape model; and

remedying error accumulation during tracking of the one or more control points by using a multi-template adaptive matching framework.

23. The computer readable program storage device of claim 22 wherein the control points are defined manually.

24. The computer readable program storage device of claim 22 wherein the control points are defined automatically.

25. The computer readable program storage device of claim 22 wherein tracking each of the one or more control points employs a Bayesian kernel matching approach.

26. The computer readable program storage device of claim 22 wherein tracking each of the one or more control points employs an optical flow based approach.

27. The computer readable program storage device of claim 22 wherein the subspace constraint model is a Gaussian model.

28. The computer readable program storage device of claim 22 wherein the subspace shape constraint model is an eigenspace model.

29. The computer readable program storage device of claim 22 wherein the object is a myocardial wall of a left ventricle.

30. The computer readable program storage device of claim 29 wherein the myocardial wall is imaged via an echocardiogram.

31. The computer readable program storage device of claim 29 wherein the shape tracking is used to detect abnormalities in the myocardial wall.

32. The computer readable program storage device of claim 29 wherein the myocardial wall is an endocardial wall.

33. The computer readable program storage device of claim 29 wherein the myocardial wall is an epicardial wall.

34. The computer readable program storage device of claim 22 wherein each control point is tracked independently of the other control points.

35. The computer readable program storage device of claim 22 wherein the uncertainty representation includes anisotropic uncertainty.

36. The computer readable program storage device of claim 22 wherein a strongly adapted Principal Component Analysis (SA-PCA) model is used to constrain the shape of the contour.

37. The computer readable program storage device of claim 36 wherein the initial shape is used to construct the SA-PCA model.

38. The computer readable program storage device of claim 22 wherein the global shape is defined by an ordered set of control points.

39. The computer readable program storage device of claim 22 wherein the prior shape model is derived from training data.

40. The computer readable program storage device of claim 22 wherein the shape of the object in motion is deformable.

41. The computer readable program storage device of claim 40 wherein the motion is non-rigid.

42. The computer readable program storage device of claim 22, the method further comprising the step of:

displaying each subsequent shape.

43. A computer implemented method for tracking a deformable global shape of an object in motion in a sequence of digital medical images, the method performed by the computer comprising the steps of:

providing a probabilistic subspace shape model for the deformable global shape derived from a set of training shapes;

defining one or more control points along the deformable global shape;

tracking each of the one or more control points as the object is in motion;

estimating uncertainty of a location of a control point in motion, wherein estimating uncertainty includes measuring anisotropic uncertainty;

building multiple appearance models for each control point, the multiple appearance models being based on uncertainty of a location of one or more of the control points;

using robust information fusion to compute an estimate of the location and uncertainty of each control point; and

combining the appearance models, the uncertainty, and the probabilistic subspace shape model to determine a current location of the object wherein the tracked deformable global shape resides inside the shape model subspace and is a most plausible shape according to uncertainties in global motion estimation local point measurements, and the deformable shape model.

44. The method of claim 43 wherein the robust information fusion is Variable-Bandwidth Density-based Fusion (VBDF).

45. The method of claim 44 wherein robust information fusion comprises the steps of:

inputting a number of measurements represented by mean vectors and covariance matrices;

fusing together the measurements by computing variable bandwidth density estimates; and

detecting a most significant mode of the variable bandwidth density estimate.

46. The method of claim 43 wherein the object is a myocardial wall of a left ventricle.

47. The method of claim 46 wherein the myocardial wall is imaged via an echocardiogram.

48. The method of claim 46 wherein the shape tracking is used to detect abnormalities in the myocardial wall.

49. The method of claim 46 wherein the myocardial wall is an endocardial wall.

50. The method of claim 46 wherein the myocardial wall is an epicardial wall.

51. The method of claim 43 wherein each control point is tracked independently of the other control points.

52. The method of claim 43 wherein a strongly adapted Principal Component Analysis (SA-PCA) model is used to constrain the shape.

53. The method of claim 52 wherein the initial shape is used to construct the SA-PCA model.

54. The method of claim 43 wherein the global shape is defined by an ordered set of control points.

55. The method of claim 43 wherein the shape of the object in motion is deformable.

56. The method of claim 55 wherein the motion is non-rigid.