Image enhancement method and system for fiducial-less tracking of treatment targets

• US 20060002615A1
• Filed: 06/30/2004
• Published: 01/05/2006
• Est. Priority Date: 06/30/2004
• Status: Active Grant

• Current Assignee: Accuray Incorporated
• Original Assignee: Accuray Incorporated
• Inventors: Kuduvalli, Gopinath, Fu, Dongshan
• Granted Patent: US 7,522,779 B2
• Time in Patent Office: Days
• US Class Current: 382/254
• CPC Class Codes: A61B 6/505 for diagnosis of bone
• CPC Class Codes: A61B 6/5235 combining images from the same or different ionising radiation imaging techniques, e.g. PET and CT
• CPC Class Codes: A61N 2005/1061 using an x-ray imaging system having a separate imaging source
• CPC Class Codes: A61N 2005/1062 using virtual X-ray images, e.g. digitally reconstructed radiographs [DRR]
• CPC Class Codes: A61N 5/1049 for verifying the position of the patient with respect to the radiation beam
• CPC Class Codes: G06T 2207/10116 X-ray image
• CPC Class Codes: G06T 2207/30004 Biomedical image processing
• CPC Class Codes: G06T 5/30 Erosion or dilatation, e.g. thinning
• CPC Class Codes: G06T 5/94 based on local image properties, e.g. for local contrast enhancement
• CPC Class Codes: G06T 7/33 using feature-based methods
• CPC Class Codes: G06T 7/35 using statistical methods
• CPC Class Codes: G06T 7/38 Registration of image sequences

Abstract

A method and system are presented for enhancing one or more images of an object, so as to increase the visibility within the images of one or more structures within the object. The object may be an anatomical region of a patient, and may include one or more reference structures, for example skeletal structures or vertebral structures, and one or more treatment targets, for example tumors or lesions. An operator, for example a top-hat filter operator, selects at least a first neighborhood and a second neighborhood within the images. The operator selects within each neighborhood one or more pixels having an optimal pixel value, and eliminates the remaining pixels in these neighborhoods. When the operator is applied to the selected neighborhoods, only the pixels having the greatest pixel values remain in the selected neighborhoods, and the remaining pixels are eliminated in the selected neighborhoods. As a result, desired features can be located and enhanced in the images.

Claims

1. A method of enhancing an image of an object so as to increase the visibility in the image of at least one structure within the object, wherein the image is characterized by a plurality of pixels, each pixel having an associated pixel value that represents the image intensity of a corresponding unit volume of the object, the method comprising:

a) selecting at least a first neighborhood and a second neighborhood within the image,

b) constructing an operator configured to select, within the first and second neighborhoods, one or more pixels having an optimal pixel value, and to eliminate the remaining pixels in the neighborhoods; and

c) applying the operator to the selected neighborhoods so that only the pixels having the optimal pixel values remain in the selected neighborhoods, and the remaining pixels in the selected neighborhoods are eliminated.

2. A method in accordance with claim 1, wherein the first and the second neighborhoods are characterized by different sizes.

3. A method in accordance with claim 1, wherein the operator comprises a non-linear operator.

4. A method in accordance with claim 3, wherein the operator comprises a non-linear top-hat filter operator.

5. A method in accordance with claim 1, wherein the act of selecting the neighborhoods comprises:

selecting the at least two different neighborhoods within the image so that at least a portion of the structure within the object is included in at least one of the neighborhoods.

6. A method in accordance with claim 2, further comprising the act of constructing the operator so that the operator is configured to perform the following acts:

1) select, from each neighborhood, the pixel having the largest pixel value;

2) compare the pixel selected from the smaller of the first and the second neighborhood with the pixel selected from the larger of the first and the second neighborhood; and

3) if the pixel selected from the smaller neighborhood is greater than the pixel selected from the larger neighborhood by a threshold amount, keep the pixel;

otherwise, eliminate the pixel.

7. A method in accordance with claim 4, wherein a mathematical expression for the top-hat filter operator comprises:

f_e=f+w×

[f−

γ

_B(f)]−

b×

[φ

_B(f)−

f],

wherein;

f represents the image,

f_e represents an enhanced image resulting from the application of the top-hat filter to selected pixels within the image;

w and b represent weighting coefficients,

γ

_B(f) represents a first structural element for the opening of the image f, and

φ

_B(f) represents a second structural element for the closing of the image f.

8. A method in accordance with claim 4, wherein the top-hat filter operator comprises a white top-hat filter operator multiplied by a weighting coefficient, and a black top-hat filter operator multiplied by another weighting coefficient.

9. A method in accordance with claim 8,

wherein the white top-hat filter is configured to subtract a first structural element from the image, and is defined as f−

γ

_B(f), where f represents the image and γ

_B(f) represents the first structural element; and

wherein the black top-hat filter is configured to subtract the image from a second structural element, and is defined as φ

_B(f)−

f, where f represents the image and φ

_B(f) represents the second structural element.

10. A method in accordance with claim 9,

wherein at least one of the first structural element and the second structural element comprises a mask for performing a morphological operation on the image f.

11. A method in accordance with claim 7, wherein the act of constructing the operator comprises:

determining empirical values for the first structural element γ

_B(f) and the second structural element φ

_B(f); and

adaptively determining the values of the weighting coefficients w and b using the amplitudes of the first and second structural elements.

12. A method in accordance with claim 1, wherein the object comprises an anatomical region of a patient, and the structure comprises at least one skeletal structure and at least one treatment target.

13. A method in accordance with claim 12, wherein the anatomical region comprises at least one of:

a cervical region of a spine;

a lumbar region of the spine; and

a thoracic region of the spine; and

wherein the reference structure comprises a skeletal structure.

14. A method in accordance with claim 12, wherein the anatomical region comprises a cervical region of a spine, and wherein the value of the weighting coefficient w is about 1, and the value of the weighting coefficient b is about 1.

15. A method in accordance with claim 12, wherein the anatomical region comprises a lumbar region of a spine, and wherein the value of the weighting coefficient w is greater than 2, and the value of the weighting coefficient b greater than 2.

16. A method of enhancing a first image and a second image of an object when registering the first image with the second image, so as to increase the visibility in the first and second images of at least one structure within the object, wherein each image is characterized by an array of pixels, each pixel having an associated pixel value that represents the image intensity of a corresponding unit volume of the object, the method comprising:

a) constructing an operator that is configured as a weighted combination of an opening of each image with a first structural element, and a closing of each image with a second structural element; and

b) applying the operator to selected pixels of each image.

17. A method in accordance with claim 16, wherein the operator comprises a non linear top-hat filter operator configured to select, from at least two different neighborhoods within each image, the pixel having an optimum associated pixel value, and to eliminate the remaining pixels in the neighborhoods.

18. A method in accordance with claim 17, wherein a mathematical expression for the top-hat filter operator comprises:

f_e=f+w×

[f−

γ

_B(f)]b×

[φ

_B(f)−

f],

wherein;

f represents the first image or the second image,

f_e represents an enhanced image resulting from the application of the top-hat filter to selected pixels within the image f;

w and b represent weighting coefficients;

γ

_B(f) represents the first structural element for combining with the opening of the image f, and

φ

_B(f) represents the second structural element for combining with the closing of the image f.

19. A method in accordance with claim 18, wherein the object comprises an anatomical region of a patient, and the structure comprises at least one reference structure and at least one treatment target.

20. A method in accordance with claim 19, wherein the anatomical region comprises at least one of:

a cervical region of a spine;

a lumbar region of the spine; and

a thoracic region of the spine; and

wherein the reference structure comprises a skeletal structure.

21. A method in accordance with claim 18, wherein the top-hat filter operator comprises a white top-hat filter operator multiplied by a weighing coefficient, and further comprises a black top-hat filter operator multiplied by another weighting coefficient;

wherein the white top-hat filter is configured to subtract the first structural element from the image f, and is defined as f−

γ

_B(f), where γ

_B(f) represents the first structural element; and

wherein the black top-hat filter is configured to subtract the image f from the second structural element, and is defined as φ

_B(f)−

f, where φ

_B(f) represents the second structural element

22. A method in accordance with claim 16, wherein the first image comprises a DRR (digitally reconstructed radiograph) that has been re-constructed from 3D scan data of the object, wherein the second image comprises an x-ray projection image of the object, and wherein the object comprises an anatomical region of a patient.

23. An apparatus for enhancing an image of an object so as to increase the visibility in the image of at least one structure within the object, the image being characterized by a plurality of pixels, each pixel having an associated pixel value that represents the image intensity of a corresponding unit volume of the object, the apparatus comprising:

a) a neighborhood selector configured to select at least a first neighborhood and a second neighborhood within the image;

b) an operator generator for constructing an operator configured to select, within the first and second neighborhoods, one or more pixels having an optimal pixel value, and to eliminate the remaining pixels in the selected neighborhoods; and

c) a controller for applying the operator to the selected neighborhoods so that only the pixels having the optimal pixel values remain in the selected neighborhoods, and the remaining pixels in the selected neighborhoods are eliminated.

24. An apparatus in accordance with claim 23, wherein the operator comprises a non linear operator.

25. An apparatus in accordance with claim 23,

wherein the operator comprises a non linear top-hat filter operator, the top-hat filter operator including a first structural element for an opening of the image, and a second structural element for a closing of the image;

and

wherein the top-hat filter operator is configured as a weighted combination of the opening of the image with the first structural element, and a closing of the image with the second structural element.

26. An apparatus in accordance with claim 23, wherein the neighborhood selector is configured to select the first and second neighborhoods within the image so that at least a portion of the structure is included in at least one of the selected neighborhoods.

27. An apparatus in accordance with claim 25, wherein a mathematical operation for the operator comprises:

f_e=f+w×

[f−

γ

_B(f)]−

b×

[φ

_B(f)−

f],

wherein;

f represents the image,

f_e represents an enhanced image resulting from the application of the operator to one or more pixels within the image;

w and b represent weighting coefficients,

γ

_B(f) represents the first structural element for the opening of the image f, and

φ

_B(f) represents the second structural element for the closing of the image f.

28. An apparatus in accordance with claim 25, wherein the top-hat filter operator comprises a white top-hat filter operator multiplied by a weighing coefficient, and a black top-hat filter operator multiplied by another weighting coefficient.

29. An apparatus in accordance with claim 28,

wherein the white top-hat filter is configured to subtract the first structural element from the image f, and is defined as f−

γ

_B(f), where γ

_B(f) represents the first structural element; and

wherein the black top-hat filter is configured to subtract the image f from the second structural element, and is defined as φ

_B(f)−

f, where φ

_B(f) represents the second structural element

30. An apparatus in accordance with claim 25,

wherein at least one of the first structural element and the second structural element comprises a mask for performing a morphological operation on the image f.

31. An apparatus in accordance with claim 27,

wherein the first structural element γ

_B(f) and the second structural element φ

_B(f) are determined empirically, and

wherein the weighting coefficients w and b are determined adaptively using the amplitudes of the empirically determined first and second structural elements.

32. An apparatus in accordance with claim 23, wherein the object comprises an anatomical region of a patient, and the structure comprises at least one reference structure and at least one treatment target.

33. An apparatus in accordance with claim 32, wherein the anatomical region comprises at least one of:

a cervical region of a spine;

a lumbar region of the spine;

and a thoracic region of the spine; and

wherein the reference structure comprises a skeletal structure.

34. An apparatus in accordance with claim 33, wherein the anatomical region comprises a cervical region of a spine, and wherein the value of the weighting coefficient w is about 1, and the value of the weighting coefficient b is about 1.

35. An apparatus in accordance with claim 33, wherein the anatomical region comprises a lumbar region of a spine, and wherein the value of the weighting coefficient w is greater than 2, and the value of the weighting coefficient b greater than 2.

36. An apparatus in accordance with claim 23, wherein the first and second neighborhoods are characterized by different sizes.

37. An apparatus in accordance with claim 36,

wherein the operator is configured to select, from each neighborhood, the pixel having the largest pixel value;

wherein the operator is further configured to compare the pixel selected from the smaller of the first neighborhood and the second neighborhood with the pixel selected from the larger of the first neighborhood and the second neighborhood; and

wherein the operator is further configured to keep the selected pixel if the pixel selected from the smaller neighborhood is greater than the pixel selected from the larger neighborhood by a threshold amount, and to otherwise eliminate the selected pixel.

38. An apparatus for enhancing a first image and a second image of an object when registering the first image with the second image, so as to increase the visibility in the first and second images of at least one structure within the object, wherein each image is characterized by an array of pixels, each pixel having an associated pixel value that represents the image intensity of a corresponding unit volume of the object, the apparatus comprising:

a) an operator generator configured to construct an operator that is configured as a weighted combination of an opening of each image with a first structural element, and a closing of each image with a second structural element; and

b) a controller configured to apply the operator to selected pixels of each image.

39. An apparatus in accordance with claim 38, wherein the operator comprises a non linear top-hat filter operator configured to select, from at least two different neighborhoods within the first image, the pixel having an optimum associated pixel value, and to eliminate the remaining pixels in the neighborhood.

40. An apparatus in accordance with claim 39, wherein a mathematical expression for the top-hat filter operator comprises:

f_e=f+w×

[f−

γ

_B(f)]−

b×

[φ

_B(f)−

f],

wherein;

f represents the first image or the second image,

f_e represents an enhanced image resulting from the application of the top-hat filter to selected pixels within the image f;

w and b represent weighting coefficients;

γ

_B(f) represents the first structural element for combining with the opening of the image f, and

φ

_B(f) represents the second structural element for combining with the closing of the image f.

41. An apparatus in accordance with claim 38, wherein the object comprises an anatomical region of a patient, and the structure comprises at least one reference structure and at least one treatment target.

42. An apparatus in accordance with claim 41, wherein the anatomical region comprises at least one of:

a cervical region of a spine;

a lumbar region of the spine;

and a thoracic region of the spine; and

wherein the reference structure comprises a skeletal structure.

43. An apparatus in accordance with claim 38, wherein the top-hat filter operator comprises a white top-hat filter operator multiplied by a weighing coefficient, and further comprises a black top-hat filter operator multiplied by another weighting coefficient;

wherein the white top-hat filter is configured to subtract the first structural element from each image, and is defined as f−

γ

_B(f), where f represents the first image or the second image, and γ

_B(f) represents the first structural element; and

wherein the black top-hat filter is configured to subtract each image from the second structural element, and is defined as φ

_B(f)−

f, where f represents the first image or the second image, and φ

_B(f) represents the second structural element

44. An apparatus in accordance with claim 38, wherein the first image comprises a DRR (digitally reconstructed radiograph) that has been re-constructed from 3D scan data of the object, wherein the second image comprises an x-ray projection image of the object, and wherein the object comprises an anatomical region of a patient.

45. An apparatus for enhancing an image of an object so as to increase the visibility in the image of at least one structure within the object, the image being characterized by a plurality of pixels, each pixel having an associated pixel value that represents the image intensity of a corresponding unit volume of the object, the apparatus comprising:

a) means for selecting at least a first neighborhood and a second neighborhood within the image;

b) means for constructing an operator that selects, within the first and second neighborhoods, one or more pixels having an optimal pixel value, and that eliminates the remaining pixels in the selected neighborhoods; and

c) control means for applying the operator to the selected neighborhoods so that only the pixels having the optimal pixel values remain in the selected neighborhoods, and the remaining pixels in the selected neighborhoods are eliminated.

46. A method of registering a near real-time 2D x-ray image of an anatomical region with 3D scan data representative of a preoperative image of the anatomical region, the anatomical region including at least one reference structure and at least one treatment target, the method comprising:

modifying the 3D scan data, and reconstructing from the modified 3D scan data at least one DRR;

enhancing the DRR and the 2D x-ray image so as to increase the visibility in the DRR and the 2D x-ray image of one of the reference structure and the treatment target;

generating a 3D motion field by estimating one or more 2D local motion fields within the DRR, and constructing a full 3D motion field from the local motion fields; and

determining from the full 3D motion field a set of non-rigid transformation parameters that represent the difference in the position and orientation of the reference structure and the treatment target, as shown in the 2D x-ray image, as compared to the position and orientation of the reference structure and the treatment target, as shown in the DRR.

47. A method in accordance with claim 46,

wherein the anatomical region comprises bone and tissue;

wherein the 3D scan data are modified to compensate for a difference between the ratio of bone-to-tissue attenuation at the energy level of the 3D scan, and the ratio of bone-to-tissue attenuation at the energy level of the x-ray imaging beam used for the near real-time x-ray image;

wherein the act of enhancing the DRR and the 2D x-ray image comprises applying a filter operator to each image, and

wherein the filter operator comprises a top-hat filter configured to select the pixel having the optimal pixel value from each of at least two different neighborhoods within the image, and to eliminate the remaining pixels in the neighborhoods, thereby transforming each image into an enhanced image.

48. An image registration system for registering at least one 2D image of an anatomical region with previously generated 3D scan data of the anatomical region, the anatomical region including at least one treatment target and at least one reference structure, wherein the 2D image is generated in near real time by detecting one or more radiographic imaging beams after the imaging beams have traversed at least a portion of the anatomical region, the imaging beams having known intensities and known positions and angles relative to the anatomical region, the system comprising:

means for providing the 3D scan data of the anatomical region;

a scan data modifier configured to modify the 3D scan data so as to compensate for a difference between the ratio of bone-to-tissue attenuation at the energy level of the 3D scan, and the ratio of bone-to-tissue attenuation at the energy level of the imaging beam used for the near real-time 2D image;

a DRR generator configured to generate at least one DRR (digitally reconstructed radiograph) of the anatomical region, using the 3D scan data and the known locations, angles, and intensities of the imaging beams;

an image enhancer configured to enhance the DRR and the 2D image by applying a filter operator to the DRR and to the 2D image;

a motion field generator configured to generate a 3D full motion field within the DRR by estimating a plurality of local motion fields within the DRR; and

a parameter determiner configured to determine from the 3D full motion field a set of non-rigid transformation parameters that represent the difference in the position and orientation of the treatment target as shown in the 2D image, as compared to the position and orientation of the treatment target as shown in the DRR.

49. A system in accordance with claim 48,

wherein the 3D scan data comprise a plurality of CT numbers representing the image intensity of corresponding 3D CT voxels, each CT voxel representing a corresponding volume element of the anatomical region;

wherein each CT voxel is disposed within one of a plurality of axial voxel slices, each axial voxel slice representing a corresponding axial slice of the anatomical region;

wherein each CT number represents the attenuated intensity of an x-ray CT beam that has been generated at a CT scan energy level and that has traversed the corresponding volume element of the anatomical region

50. A system in accordance with claim 49,

wherein the scan data modifier includes a processor for performing on each CT number a mathematical operation derived from a non-linear x-ray attenuation model;

and

wherein the mathematical operation is given by;

C(x,y,z)=aC₀(x,y,z)e^bC0(x,y,z) 

 

(1)

where C(x,y,z) represents the modified CT number of a 3D CT voxel having a location (x,y,z);

a and b represent weighting coefficients;

and C₀(x,y,z) represents the un-modified CT number, based on a linear attenuation model, of a 3D CT voxel having a location (x,y,z).

51. A system in accordance with claim 48,

wherein the image enhancer comprises:

a) an operator generator configured to construct an operator that is configured as a weighted combination of an opening of each image with a first structural element, and a closing of each image with a second structural element; and

b) a controller configured to apply the operator to selected pixels of each image.

52. A system in accordance with claim 51, wherein the operator comprises a non linear top-hat filter operator configured to select, from at least two different neighborhoods within the first image, the pixel having an optimum associated pixel value, and to eliminate the remaining pixels in the neighborhood.

53. A system in accordance with claim 52, wherein a mathematical expression for the top-hat filter operator comprises:

f_e=f+w×

[f−

γ

_B(f)]−

b×

[φ

_B(f)−

f],

wherein;

f represents the first image or the second image,

f_e represents an enhanced image resulting from the application of the top-hat filter to selected pixels within the image f;

w and b represent weighting coefficients;

γ

_B(f) represents the first structural element for combining with the opening of the image f, and

φ

_B(f) represents the second structural element for combining with the closing of the image f.

54. A system in accordance with claim 48, wherein the motion field generator comprises:

a mesh grid generator configured to generate in the first image a mesh grid having a plurality of mesh nodes;

a nodal motion estimator configured to estimate, for each mesh node in the first image, at least one nodal motion vector that describes a matching of the mesh node with a corresponding mesh node in the second image; and

a motion field interpolator configured to determine a local motion vector for each of a plurality of points of interest within the first image, by interpolating from the nodal motion vectors of the mesh nodes that surround each point of interest.

55. A system in accordance with claim 54,

wherein the mesh grid generator is configured to repeat, at each of a plurality of mesh resolution levels, the act of generating of the mesh grid;

wherein nodal motion estimator is configured to repeat, at each of the plurality of mesh resolution levels, the act of estimating the at least one nodal motion vector for each mesh node; and

wherein the nodal motion vectors from which the motion field interpolator interpolates to determine the local motion vector comprise nodal motion vectors that have been estimated at a final one of the plurality of mesh resolution levels.

56. A system in accordance with claim 54,

further comprising a motion field reconstructor configured to reconstruct the nodal motion vector for a mesh node if any nodal motion vector estimated for any mesh node is unreliable.

57. A system in accordance with claim 55,

wherein at each mesh resolution level, the nodal motion estimator is configured to pass onto a subsequent mesh resolution level one or more nodal motion vectors that have been estimated at the current mesh resolution level.

58. A system in accordance with claim 57,

wherein at each mesh resolution level subsequent to the first mesh resolution level, the nodal motion estimator is configured to use the nodal motion vectors passed on from a previous mesh resolution level as estimates for nodal motion vectors for the first subset of mesh nodes;

wherein the nodal motion estimator comprises an interpolator configured to interpolate from the nodal motion vectors estimated for the first subset of mesh nodes, to estimate the nodal motion vectors for the second subset of mesh nodes; and

wherein the nodal motion estimator further comprises a nodal motion refiner configured to refine, for each mesh node in both the first and the second subsets, the nodal motion vector that was estimated for the mesh node.

59. A system in accordance with claim 58,

wherein the nodal vector refiner is configured to:

define a block centered on a mesh node in the first image;

search for a matching mesh node in the second image that maximizes a similarity measure between the block in the first image and another block centered around the matching mesh node in the second image; and

modify the nodal motion vector so that the nodal motion vector describes a mapping of the mesh node in the first image onto the matching mesh node in the second image.

60. A system in accordance with claim 59,

wherein the nodal vector refiner is configured to repeat, for each of a plurality of successively increasing image resolution levels, the acts of defining a block centered around a mesh node in the first image, searching for a matching mesh node in the second image, and refining the nodal motion vector.

61. A method of registering a near real-time 2D x-ray image of an anatomical region with 3D scan data representative of a preoperative image of the anatomical region, the anatomical region including at least one reference structure and at least one treatment target, the method comprising:

reconstructing from the 3D scan data at least one DRR;

enhancing the DRR and the 2D x-ray image so as to increase the visibility in the DRR and the 2D x-ray image of one of the reference structure and the treatment target;

selecting an ROI (region of interest) within the DRR, wherein the ROI includes at least a portion of the reference structure or the treatment target;

generating a 3D motion field by estimating one or more 2D local motion fields within the ROI, and constructing a full 3D motion field from the local motion fields; and

determining from the full 3D motion field a set of non-rigid transformation parameters that represent the difference in the position and orientation of the reference structure and the treatment target, as shown in the 2D x-ray image, as compared to the position and orientation of the reference structure and the treatment target, as shown in the DRR.

62. A method in accordance with claim 61,

wherein the act of enhancing the DRR and the 2D x-ray image comprises applying a filter operator to each image, and

wherein the filter operator comprises a top-hat filter configured to select the pixel having the optimal pixel value from each of at least two different neighborhoods within the image, and to eliminate the remaining pixels in the neighborhoods, thereby transforming each image into an enhanced image.

63. A method in accordance with claim 62, wherein the filter operator is defined mathematically by:

f_e=f+w×

[f−

γ

_B(f)]−

b×

[φ

_B(f)−

f],

where f_e represents an enhanced image,

f represents an original image,

w and b represent weighting coefficients,

γ

_B(f) represents a structural element for the opening of f, and

φ

_B(f) represents a structural element for the closing of f.

64. A method in accordance with claim 61, wherein the step of defining the ROI comprises:

a) calculating an entropy measure H of the DRR; and

b) selecting an ROI so as to maximize the entropy measure H.

65. A method in accordance with claim 64, wherein the entropy measure H is a modified Shannon entropy measure given by:

H=−

Σ

IP(I)log P(I),

where I is the value of the intensity of the image, at each pixel of the image, and P(I) is the probability of an image intensity value I occurring within the ROI.

66. A method in accordance with claim 61, wherein the act of generating the 3D motion field comprising:

generating in the 2D x-ray image a mesh grid having a plurality of mesh nodes;

for each mesh node in the 2D x-ray image, estimating at least one nodal motion vector that describes a matching of the mesh node with a corresponding mesh node in the DRR; and

determining a local motion vector for each of a plurality of points of interest within the 2D x-ray image, by interpolating from the nodal motion vectors of the mesh nodes that surround the point of interest.

67. An image registration system for registering at least one 2D image of an anatomical region with previously generated 3D scan data of the anatomical region, the anatomical region including at least one treatment target and at least one reference structure, wherein the 2D image is generated in near real time by detecting one or more radiographic imaging beams after the imaging beams have traversed at least a portion of the anatomical region, the imaging beams having known intensities and known positions and angles relative to the anatomical region, the system comprising:

means for providing the 3D scan data of the anatomical region;

a DRR generator configured to generate at least one DRR (digitally reconstructed radiograph) of the anatomical region, using the 3D scan data and the known locations, angles, and intensities of the imaging beams;

an image enhancer configured to enhance the DRR and the 2D image by applying a filter operator to the DRR and to the 2D image;

an ROI selector configured to select an ROI (region of interest) within the DRR, the ROI including at least a portion of the treatment target or of the reference structure;

a motion field generator configured to generate a 3D full motion field within the DRR by estimating a plurality of local motion fields within the DRR; and

a parameter determiner configured to determine from the 3D full motion field a set of non-rigid transformation parameters that represent the difference in the position and orientation of the treatment target as shown in the 2D image, as compared to the position and orientation of the treatment target as shown in the DRR.

68. A system in accordance with claim 67,

wherein the filter operator comprises a top-hat filter configured to select the pixel having the greatest pixel value from each of at least two different neighborhoods within the DRR and the 2D image, and to eliminate the remaining pixels in the neighborhoods.

69. A system in accordance with claim 68,

wherein the filter operator is defined mathematically by:

f_e=f+w×

[f−

γ

_B(f)]−

b×

[φ

_B(f)−

f],

where f_e represents an enhanced image,

f represents an original image,

w and b represent weighting coefficients,

γ

_B(f) represents a structural element for the opening of the original image f, and

φ

_B(f) represents a structural element for the closing of the original image f.

70. A system in accordance with claim 67,

wherein the 2D image and the DRR are each characterized by an array of pixels, each pixel having an associated pixel value that represents the image intensity of a corresponding unit of the anatomical region.

71. A system in accordance with claim 67,

wherein the ROI selector comprises an entropy calculator for calculating an entropy measure H of the DRR.

72. A system in accordance with claim 71,

wherein the entropy measure H comprises a modified Shannon entropy measure given by:

H=−

Σ

IP(I)log P(I),

where I is the value of the image intensity level, and P(I) is the probability of an image intensity value I occurring within the ROI.

73. A system in accordance with claim 71,

wherein the ROI is selected so that the entropy measure H is maximized within the ROI.

74. A system in accordance with claim 67, wherein the motion field generator comprises:

a mesh grid generator configured to generate in the 2D x-ray image a mesh grid having a plurality of mesh nodes;

a nodal motion estimator configured to estimate, for each mesh node in the 2D x-ray image, at least one nodal motion vector that describes a matching of the mesh node with a corresponding mesh node in the DRR; and

a motion field interpolator configured to determine a local motion vector for each of a plurality of points of interest within the 2D x-ray image, by interpolating from the nodal motion vectors of the mesh nodes that surround each point of interest.

75. A system in accordance with claim 74,

wherein the mesh grid generator is configured to repeat, at each of a plurality of mesh resolution levels, the act of generating of said mesh grid;

wherein nodal motion estimator is configured to repeat, at each of said plurality of mesh resolution levels, the act of estimating the at least one nodal motion vector for each mesh node; and

wherein the nodal motion vectors from which the motion field interpolator interpolates to determine the local motion vector comprise nodal motion vectors that have been estimated at a final one of the plurality of mesh resolution levels.

76. A system in accordance with claim 74,

wherein the motion field generator further comprises a motion field reconstructor configured to reconstruct the nodal motion vector for a mesh node if any nodal motion vector estimated for any mesh node is unreliable.

77. An image registration system for registering at least one 2D image of an anatomical region with previously generated 3D scan data of the anatomical region, the anatomical region including at least one treatment target and at least one reference structure, wherein the 2D image is generated in near real time by detecting one or more radiographic imaging beams after the imaging beams have traversed at least a portion of the anatomical region, the imaging beams having known intensities and known positions and angles relative to the anatomical region, the system comprising:

means for providing the 3D scan data of the anatomical region;

a DRR generator configured to generate at least one DRR (digitally reconstructed radiograph) of the anatomical region, using the 3D scan data and the known locations, angles, and intensities of the imaging beams;

an image enhancer configured to enhance the DRR and the 2D image by applying a filter operator to the DRR and to the 2D image;

a motion field generator configured to generate a 3D full motion field within the DRR by estimating a plurality of local motion fields within the DRR; and

a parameter determiner configured to determine from the 3D full motion field a set of non-rigid transformation parameters that represent the difference in the position and orientation of the treatment target as shown in the 2D image, as compared to the position and orientation of the treatment target as shown in the DRR.

78. A system in accordance with claim 77, wherein the motion field generator comprises:

a mesh grid generator configured to generate in the 2D x-ray image a mesh grid having a plurality of mesh nodes;

a nodal motion estimator configured to estimate, for each mesh node in the 2D x-ray image, at least one nodal motion vector that describes a matching of the mesh node with a corresponding mesh node in the DRR; and

a motion field interpolator configured to determine a local motion vector for each of a plurality of points of interest within the 2D x-ray image, by interpolating from the nodal motion vectors of the mesh nodes that surround each point of interest.