Detection and analysis of lesions in contact with a structural boundary
View Patent ↗An algorithm is disclosed that detects and analyzes possible lesions that are left in contact with a structural boundary. A morphological closing with a structuring element is performed along the boundary to detect lesions within the threshold of the structuring element. A deformable surface-based analysis is performed on distinctive surfaces of the structure for the identification of lesions. The integrated use of a deformable surface model and chamfer distance potential enables explicit representation of regularized, or smoothed, surfaces from which lesion candidates may be detected.
1. A method of recovering from digital images lesions in contact with a boundary of an anatomical structure comprising:
performing a morphological operation at the boundary of the anatomical structure to identify lesions smaller than a first threshold size; and
performing a surface-regularization operation on at least one surface of the structure to identify lesions.
2. The method of claim 1 wherein the surface-regularization operation detects lesions larger than a second threshold size.
3. The method of claim 2 wherein the first and second threshold sizes are the same.
4. The method of claim 2 wherein the second threshold size is larger than the first threshold size.
5. The method of claim 2 wherein the second threshold size is smaller than the first threshold size.
6. The method of claim 1 wherein the anatomical structure is a lung field.
7. The method of claim 1 wherein the anatomical structure is a heart, brain, spine, colon, liver or kidney.
8. The method of claim 1 wherein the boundary at which the morphological operation is performed is a boundary of a digital volume constructed from the digital images.
9. The method of claim 8 wherein the boundary at which the morphological operation is performed is selected after segmenting the digital volume.
10. The method of claim 9 wherein segmenting the digital volume further includes identifying a lung field region.
11. The method of claim 9 wherein segmenting the digital volume further includes identifying a pleural boundary.
12. The method of claim 9 wherein segmenting the digital volume further includes identifying a lobe of a lung.
13. The method of claim 1 wherein the boundary at which the morphological operation is performed is selected after segmenting one or more of the digital images.
14. The method of claim 1 further including segmenting one or more of the digital images to identify surfaces for surface fitting and regularization in smoothness.
15. The method of claim 1 further including segmenting a digital volume created from the digital images to identify surfaces on which to perform the surface-regularization operation.
16. The method of claim 15 wherein the surface-regularization operation is performed on a lung surface.
17. The method of claim 16 wherein the lung surface is a lung apex, a costal surface and posterior border, a lung base, or a mediastinal surface and anterior border.
18. The method of claim 1 wherein the surface-regularization operation includes surface-fitting and regularization in smoothness with deformable models.
19. The method of claim 18 wherein the surface-fitting with deformable models is initialized using landmark points from a section of one or more digital images.
20. The method of claim 19 wherein the landmark points identify a lung apex, a costal surface and posterior border, a lung base, or a mediastinal surface and anterior border.
21. The method of claim 1 wherein performing a morphological operation at the boundary of the anatomical structure further includes:
performing a morphological dilation at the boundary with a structuring element; and
performing morphological erosion at the boundary with the structuring element.
22. The method of claim 21 wherein the structuring element is ellipsoidal.
23. The method of claim 21 wherein the structuring element is spherical.
24. The method of claim 23 wherein the spherical structuring element has a radius of approximately 3 millimeters.
25. The method of claim 23 wherein the structuring element has a radius of 2 to 4 millimeters.
26. The method of claim 23 wherein the first threshold size is the radius of the structuring element.
27. The method of claim 1 wherein the first threshold size is 3 millimeters.
28. The method of claim 1 further including displaying one of the digital images, said digital image having a location associated with an area corresponding to a location identified by the morphological operation or the surface-regularization operation.
29. The method of claim 28 wherein the displayed image includes one or more highlighted abnormalities.
30. The method of claim 29 wherein displaying a digital image further includes displaying at least one of the digital images of the anatomical structure, said at least one of the digital images of the anatomical structure being a digital image not processed by the morphological operation or surface-regularization operation and containing anatomical information associated with anatomical information of the displayed digital image.
31. The method of claim 30 wherein the digital image of the anatomical structure displayed is shown with a highlighted location, said highlighted location associated with one or more of the highlighted abnormalities in the displayed digital image.
32. The method of claim 1 further including displaying one of the digital images, said digital image having a location associated with an area corresponding to a location identified by the morphological operation or the surface-regularization operation.
33. The method of claim 1 further including displaying a post-processed digital image volume, said post-processed digital image volume constructed from digital images identified as containing anatomical information identified by the morphological operation or the surface-regularization operation.
34. The method of claim 1 further including the step of performing noise reduction on the digital images.
35. The method of claim 1 further including the step of performing noise reduction on a digital volume constructed from some or all of the digital images.
36. The method of claim 1 wherein the step of performing a morphological operation is performed on a digital lung field image.
37. The method of claim 1 wherein an abnormality identified by the morphological operation or the surface-regularization operation is a pleural nodule.
38. The method of claim 1 wherein a surface-regularization operation is performed using a deformable surface model to fit a surface to a boundary.
39. The method of claim 38 wherein the deformable surface model uses a chamfer distance potential as a measure of the proximity of the deformable surface model to the boundary.
40. The method of claim 1 wherein the surface of the structure upon which surface-regularization operation is performed is a pleural surface.
41. The method of claim 1 wherein performing a surface-regularization operation requires fitting a pleural surface to a boundary.
42. The method of claim 41 wherein the boundary is a pleural boundary.
43. The method of claim 1 wherein performing a surface-regularization operation on at least one surface further includes:
identifying border voxels;
calculating a potential to the border voxels;
initializing the surface; and
deforming a deformable surface model under an external force and under an internal force.
44. The method of claim 43 wherein the potential is a chamfer distance potential.
45. The method of claim 43 wherein the external force is based on a chamfer distance potential.
46. The method of claim 43 wherein the internal force is based on a smoothness determination of the surface.
47. The method of claim 43 wherein the surface includes a costal surface and a posterior border of the lungs.
48. The method of claim 43 wherein the surface is a base of the lungs.
49. The method of claim 43 wherein the surface includes a mediastinal surface of the lungs and an anterior border of the lungs.
50. The method of claim 43 wherein the surface includes a lung apex.
51. The method of claim 1 further including highlighting abnormalities in the digital images corresponding to abnormalities identified by the morphological operation or by the surface-regularization operation.
52. The method of claim 1 further including highlighting abnormalities in a digital volume constructed from digital images corresponding to abnormalities identified by the morphological operation or by the surface-regularization operation.
53. The method of claim 1 wherein the lesion is a protrusion from the boundary of the anatomical structure.
54. The method of claim 1 wherein the lesion is an indentation in the boundary of the anatomical structure.
55. The method of claim 1 wherein the digital images are formed by computer tomography, magnetic resonance imaging, positron emission tomography or single photon emission computer tomography.
56. A method of recovering from digital images lesions in contact with a boundary of an anatomical structure comprising:
recovering lesions having a size below a first threshold by performing a morphological operation at the boundary of the anatomical structure comprising:
performing a morphological dilation at the boundary with a structuring element selected to be capable of detecting lesions having a size below the first threshold; and
performing morphological erosion at the boundary with the structuring element; and
performing a surface-regularization operation on at least one surface of the structure to identify lesions comprising:
identifying border voxels;
calculating a potential to the border voxels;
initializing the surface; and
deforming a deformable surface model under an external force and under an internal force.
57. The method of claim 56 wherein the potential is a chamfer distance potential.
58. The method of claim 56 wherein the external force is based on a chamfer distance potential.
59. The method of claim 56 wherein the internal force is based on a smoothness determination of the surface.
60. The method of claim 56 wherein the surface is a lung apex, a costal surface and posterior border, a lung base, or a mediastinal surface and anterior border.
61. A method of recovering from digital image pleural nodules comprising:
performing a morphological operation on pleural surfaces to identify nodules smaller than a first size; and
performing a surface-regularization operation on at least one pleural surface to identify other nodules.
62. In a recording medium, software for recovering from digital images lesions in contact with a boundary of an anatomical structure comprising:
logic code for performing a morphological operation at the boundary of the anatomical structure to identify lesions having a size smaller than a first threshold; and
logic code for performing a surface-regularization operation on at least one surface of the structure to identify abnormalities.