System and method for airway detection

1. A method for detecting airways in 3-dimensional image data, comprising:

receiving 3D lung image data;

filtering, by a processor, the 3D lung image data using one or more filters based on multiple first and second order derivatives of the 3D lung image data; and

detecting, by a processor, airways in the 3D lung image data based on the filtered 3D lung image data, wherein the 3D lung image data comprises a plurality of voxels, and said step of filtering the 3D lung image data comprises:

calculating a gradient at each voxel of said plurality of voxels;

filtering each voxel of the 3D lung image data using a first filter based on the gradient calculated at each voxel to calculate a first filter value for each voxel;

calculating a curvature measure at each voxel of said plurality of voxels;

filtering each voxel of the 3D lung image data using a second filter based on the curvature calculated at each voxel to calculate a second filter value for each voxel; and

combining the first and second filter values for each voxel to determine a voxel score associated with each voxel.

2. The method of claim 1 , wherein said step of detecting airways comprises:

comparing the voxel score associated with each voxel to a threshold value; and

determining voxels for which the voxel score is greater than or equal to the threshold value to be airway candidates.

3. The method of claim 1 , wherein said first filter comprises a magnitude-angle-radius (MARS) filter.

4. The method of claim 1 , wherein said second filter is based on principal curvatures.

5. The method of claim 1 , wherein the 3D lung image data is acquired from Computed Tomography (CT) imaging.

6. A method for detecting airways in 3-dimensional image data, comprising:

receiving 3D lung image data;

filtering, by a processor, the 3D lung image data using one or more filters based on multiple first and second order derivatives of the 3D lung image data; and

detecting, by a processor, airways in the 3D lung image data based on the filtered 3D lung image data, wherein the 3D lung image data comprises a plurality of voxels, and said step of filtering the 3D lung image data comprises:

calculating a gradient at each voxel of said plurality of voxels;

filtering each voxel of the 3D lung image data using a first filter based on the gradient calculated at each voxel to calculate a first filter value for each voxel;

calculating vesselness at each voxel of said plurality of voxels;

filtering each voxel of the 3D lung image data using a second filter based on the vesselness calculated at each voxel to calculate a second filter value for each voxel; and

combining the first and second filter values for each voxel to determine a voxel score associated with each voxel.

7. The method of claim 6 , wherein said step of detecting airways comprises:

comparing the voxel score associated with each voxel to a threshold value; and

determining voxels for which the voxel score is greater than or equal to the threshold value to be airway candidates.

8. The method of claim 6 , wherein said first filter comprises a magnitude-angle-radius (MARS) filter.

9. The method of claim 6 , wherein the 3D lung image data is acquired from Computed Tomography (CT) imaging.

10. A method for detecting airways in 3-dimensional image data, comprising:

receiving 3D lung image data;

filtering, by a processor, the 3D lung image data using one or more filters based on multiple first and second order derivatives of the 3D lung image data; and

detecting, by a processor, airways in the 3D lung image data based on the filtered 3D lung image data, wherein the 3D lung image data comprises a plurality of voxels, and said step of filtering the 3D lung image data comprises:

determining a filter value using an airway filter based on the first order derivative of the 3D lung image data in the vicinity of each voxel and an airway wall filter based on the first and second order derivatives of the 3D lung image data in the vicinity of each voxel.

11. The method of claim 10 , wherein the 3D lung image data is acquired from Computed Tomography (CT) imaging.

12. An apparatus for detecting airways in 3D lung image data, comprising:

means for filtering the 3D lung image data using one or more filters based on multiple first and second order derivatives of the 3D lung image data; and

means for detecting airways in the 3D lung image data based on the filtered 3D lung image data, wherein the 3D lung image data comprises a plurality of voxels, and said means for filtering the 3D lung image data comprises:

means for calculating a gradient at each voxel of said plurality of voxels;

a first filter for filtering each voxel of the 3D lung image data based on the gradient calculated at each voxel to calculate a first filter value for each voxel;

means for calculating a curvature measure at each voxel of said plurality of voxels;

a second filter for filtering each voxel the 3D lung image data based on the curvature calculated at each voxel to calculate a second filter value for each voxel; and

means for combining the first and second filter values for each voxel to determine a voxel score associated with each voxel.

13. The apparatus of claim 12 , wherein said first filter comprises a magnitude-angle-radius (MARS) filter.

14. The apparatus of claim 12 , wherein said second filter filters each voxel based on computed principal curvatures.

15. The apparatus of claim 12 , wherein the 3D lung image data is acquired from Computed Tomography (CT) imaging.

16. An apparatus for detecting airways in 3D lung image data, comprising:

means for filtering the 3D lung image data using one or more filters based on multiple first and second order derivatives of the 3D lung image data; and

means for detecting airways in the 3D lung image data based on the filtered 3D lung image data, wherein the 3D lung image data comprises a plurality of voxels, and said means for filtering the 3D lung image data comprises:

means for calculating a gradient at each voxel of said plurality of voxels;

a first filter for filtering each voxel of the 3D lung image data based on the gradient calculated at each voxel to calculate a first filter value for each voxel;

means for calculating vesselness at each voxel of said plurality of voxels;

a second filter for filtering each voxel the 3D lung image data based on the vesselness calculated at each voxel to calculate a second filter value for each voxel; and

means for combining the first and second filter values for each voxel to determine a voxel score associated with each voxel.

17. The apparatus of claim 16 , wherein said first filter comprises a magnitude-angle-radius (MARS) filter.

18. The apparatus of claim 16 , wherein the 3D lung image data is acquired from Computed Tomography (CT) imaging.

19. An apparatus for detecting airways in 3D lung image data, comprising:

means for filtering the 3D lung image data using one or more filters based on multiple first and second order derivatives of the 3D lung image data; and

means for detecting airways in the 3D lung image data based on the filtered 3D lung image data, wherein the 3D lung image data comprises a plurality of voxels, and said means for filtering the 3D lung image data comprises:

an airway filter to determine a first filter value for each voxel based on the first order derivative of the 3D lung image data in the vicinity of each voxel;

an airway wall filter to determine a second filter value for each voxel based on the second order derivative of the 3D lung image data in the vicinity of each voxel.

20. The apparatus of claim 19 , wherein the 3D lung image data is acquired from Computed Tomography (CT) imaging.

21. A non-transitory computer readable medium storing computer program instructions for detecting airways in 3D lung image data, said computer program instructions defining the steps comprising:

filtering the 3D lung image data based on multiple first and second order derivatives of the 3D lung image data; and

detecting airways in the 3D lung image data based on the filtered 3D lung image data, wherein the 3D lung image data comprises a plurality of voxels, and the computer program instructions defining the step of filtering the 3D lung image data comprise computer program instructions defining the steps of:

calculating a gradient at each voxel of said plurality of voxels;

filtering each voxel of the 3D lung image data based on the gradient calculated at each voxel to calculate a first filter value for each voxel;

calculating a curvature measure at each voxel of said plurality of voxels;

filtering each voxel the 3D lung image data based on the curvature calculated at each voxel to calculate a second filter value for each voxel; and

combining the first and second filter values for each voxel to determine a voxel score associated with each voxel.

22. The non-transitory computer readable medium of claim 21 , wherein the computer program instructions defining the step of detecting airways comprise computer program instructions defining the step of:

comparing the voxel score associated with each voxel to a threshold value; and

determining voxels for which the voxel score is greater than or equal to the threshold value to be airway candidates.

23. The non-transitory computer readable medium of claim 21 , wherein the 3D lung image data is acquired by Computed Tomography (CT) imaging.

24. A non-transitory computer readable medium storing computer program instructions for detecting airways in 3D lung image data, said computer program instructions defining the steps comprising:

filtering the 3D lung image data based on multiple first and second order derivatives of the 3D lung image data; and

detecting airways in the 3D lung image data based on the filtered 3D lung image data, wherein the 3D lung image data comprises a plurality of voxels, and the computer program instructions defining the step of filtering the 3D lung image data comprise computer program instructions defining the steps of:

calculating a gradient at each voxel of said plurality of voxels;

filtering each voxel of the 3D lung image data based on the gradient calculated at each voxel to calculate a first filter value for each voxel;

calculating vesselness at each voxel of said plurality of voxels;

filtering each voxel the 3D lung image data based on the vesselness calculated at each voxel to calculate a second filter value for each voxel; and

combining the first and second filter values for each voxel to determine a voxel score associated with each voxel.

25. The non-transitory computer readable medium of claim 24 , wherein the computer program instructions defining the step of detecting airways comprise computer program instructions defining the step of:

comparing the voxel score associated with each voxel to a threshold value; and

determining voxels for which the voxel score is greater than or equal to the threshold value to be airway candidates.

26. The non-transitory computer readable medium of claim 24 , wherein the 3D lung image data is acquired by Computed Tomography (CT) imaging.

27. A non-transitory computer readable medium storing computer program instructions for detecting airways in 3D lung image data, said computer program instructions defining the steps comprising:

filtering the 3D lung image data based on multiple first and second order derivatives of the 3D lung image data; and

detecting airways in the 3D lung image data based on the filtered 3D lung image data, wherein the 3D lung image data comprises a plurality of voxels, and the computer program instructions defining the step of filtering the 3D lung image data comprise computer program instructions defining the step of:

determining a filter value using an airway filter based on the first order derivative of the 3D lung image data in the vicinity of each voxel and an airway wall filter based on the first and second order derivatives of the 3D lung image data in the vicinity of each voxel.

28. The non-transitory computer readable medium of claim 27 , wherein the 3D lung image data is acquired by Computed Tomography (CT) imaging.