Techniques for segmentation of lymph nodes, lung lesions and other solid or part-solid objects
Techniques for segmentation include determining an edge of voxels in a range associated with a target object. A center voxel is determined. Target size is determined based on the center voxel. In some embodiments, edges near the center are suppressed, markers are determined based on the center, and an initial boundary is determined using a watershed transform. Some embodiments include determining multiple rays originating at the center in 3D, and determining adjacent rays for each. In some embodiments, a 2D field of amplitudes is determined on a first dimension for distance along a ray and a second dimension for successive rays in order. An initial boundary is determined based on a path of minimum cost to connect each ray. In some embodiments, active contouring is performed using a novel term to refine the initial boundary. In some embodiments, boundaries of part-solid target objects are refined using Markov models.
1. A method comprising:
obtaining, on a processor, image data that indicates amplitude values at each of a plurality of voxels for a particular measurement modality;
determining, on a processor, an outer edge of a contiguous subset of voxels based on voxels that have an amplitude in a first amplitude range associated with a target object in image data of the particular measurement modality;
determining, on a processor, a center voxel, wherein a center distance from the center voxel to a closest voxel of the outer edge is greater than a distance from any neighbor of the center voxel to a corresponding closest voxel of the outer edge;
generating new image data by causing, on a processor, a location and size of the target object to be marked in the image data based on the center voxel and the center distance further comprising
determining, on a processor, a gradient value at a voxel; and
determining, on a processor, a reduced gradient equal to a product of the gradient with a fraction f that decreases with decreasing distance of the voxel from the center voxel,
wherein the new image data is based at least in part on the reduced gradient; and
presenting on a display device output image data based at least in part on the new image data with the location and size of the target object marked.
2. The method as recited in claim 1 , further comprising determining the first amplitude range associated with the target object based on a Gaussian mixture model.
3. The method as recited in claim 1 , further comprising:
determining, on a processor, a marker based on the location and size of the target object; and
determining on a processor a watershed boundary based on the marker,
wherein the new image data is based at least in part on the watershed boundary.
4. A method comprising:
obtaining, on a processor, image data that indicates amplitude values at each of a plurality of voxels for a particular measurement modality;
determining, on a processor, an outer edge of a contiguous subset of voxels based on voxels that have an amplitude in a first amplitude range associated with a target object in image data of the particular measurement modality;
determining, on a processor, a center voxel, wherein a center distance from the center voxel to a closest voxel of the outer edge is greater than a distance from any neighbor of the center voxel to a corresponding closest voxel of the outer edge;
generating new image data by causing, on a processor, a location and size of the target object to be marked in the image data based on the center voxel and the center distance further comprising
determining, on a processor, a gradient value at each voxel;
determining, on a processor, a reduced gradient equal to a product of the gradient with a fraction f that decreases with decreasing distance of each voxel from the center voxel; and
determining, on a processor, a watershed boundary based, at least in part, on the reduced gradient instead of the gradient at voxels within the center distance from the center voxel; and
presenting on a display device output image data based at least in part on the new image data with the location and size of the target object marked on the watershed boundary.
5. A method comprising:
obtaining, on a processor, image data that indicates amplitude values at each of a plurality of voxels for a particular measurement modality;
determining, on a processor, an outer edge of a contiguous subset of voxels based on voxels that have an amplitude in a first amplitude range associated with a target object in image data of the particular measurement modality;
determining, on a processor, a center voxel, wherein a center distance from the center voxel to a closest voxel of the outer edge is greater than a distance from any neighbor of the center voxel to a corresponding closest voxel of the outer edge;
determining, on a processor, a location and size of the target object in the image data based on the center voxel and the center distance further comprising determining, on a processor, a refined boundary by moving an initial boundary based on active contouring using as the initial boundary at least one of a marker as the center voxel and the center distance or a watershed boundary; and
presenting on a display device output image data based at least in part on the refined boundary,
wherein
the active contouring of the initial boundary further comprises determining a change in boundary position based, at least in part, on a potential well term of an equation comprising the potential well term and one or more other terms, and wherein the potential well term is weighted by a parameter α independently of the other terms of the equation wherein α has a value greater than 1; and
the potential well term comprises a product of a gradient of an edge indicator function g and a gradient of a level set function ϕ.
6. The method as recited in claim 5 , wherein:
the active contouring of the initial boundary further comprises determining a change in boundary position based, at least in part, on a topographical distance term, (Li-Le), wherein
topographical distance L is a minimum accumulated weighted change in amplitude,
Li is a topographical distance to a boundary based on eroding the initial boundary by a plurality of voxels, and Le is a topographical distance to a boundary based on dilating the initial boundary by a plurality of voxels.
7. A method comprising:
obtaining, on a processor, image data that indicates amplitude values at each of a plurality of voxels for a particular measurement modality;
determining, on a processor, an outer edge of a contiguous subset of voxels based on voxels that have an amplitude in a first amplitude range associated with a target object in image data of the particular measurement modality;
determining, on a processor, a center voxel, wherein a center distance from the center voxel to a closest voxel of the outer edge is greater than a distance from any neighbor of the center voxel to a corresponding closest voxel of the outer edge;
determining, on a processor, a location and size of the target object in the image data based on the center voxel and the center distance further comprising determining, on a processor, a refined boundary by moving an initial boundary based on active contouring using as the initial boundary at least one of a marker as the center voxel and the center distance or a watershed boundary; and
presenting on a display device output image data based at least in part on the location and size of the target object or the refined boundary,
wherein:
the active contouring of the initial boundary further comprises determining a change in boundary position based, at least in part, on a potential well term independently weighted by a parameter α,
the potential well term comprises a product of a gradient of an edge indicator function g and a gradient of a level set function ϕ,
the active contouring of the initial boundary further comprises determining a change in boundary position based, at least in part, on a volume preserving mean curvature term of an equation comprising the volume preserving mean curvature term and the potential well term and wherein the volume preserving mean curvature term is independently weighted by an independent parameter β such that the potential well term is not weighted by the independent parameter β; and
the volume preserving mean curvature term comprises a product of a magnitude of the gradient of the level set function ϕ and a difference between a local mean curvature κ and a mean curvature for the initial boundary κmean.
8. A method comprising:
obtaining, on a processor, image data that indicates amplitude values at each of a plurality of voxels for a particular measurement modality;
determining, on a processor, an outer edge of a contiguous subset of voxels based on voxels that have an amplitude in a first amplitude range associated with a target object in image data of the particular measurement modality;
determining, on a processor, a center voxel, wherein a center distance from the center voxel to a closest voxel of the outer edge is greater than a distance from any neighbor of the center voxel to a corresponding closest voxel of the outer edge;
determining, on a processor, a location and size of the target object in the image data based on the center voxel and the center distance further comprising determining, on a processor, a refined boundary by moving an initial boundary based on active contouring using as the initial boundary at least one of a marker as the center voxel and the center distance or a watershed boundary; and
presenting on a display device output image data based at least in part on the location and size of the target object or the refined boundary,
wherein performing active contouring of the initial boundary comprises using an evolution equation that includes a first term that indicates deviations in amplitude for all voxels inside the boundary from a statistical metric of amplitude of all voxels inside the boundary and a second term that indicates deviations in amplitude for local voxels outside the boundary from a statistical metric of amplitude of the local voxels outside the boundary, wherein the local voxels outside the boundary exclude a larger number of voxels outside the initial boundary.
9. A non-transitory computer-readable medium carrying one or more sequences of instructions, wherein execution of the one or more sequences of instructions by one or more processors causes the one or more processors to perform the steps of claim 1 .
10. A system comprising:
a medical imaging device;
a display device;
at least one processor; and
at least one memory including one or more sequences of instructions,
the at least one memory and the one or more sequences of instructions configured to, with the at least one processor, cause an apparatus to perform at least the steps of claim 1 .
11. A non-transitory computer-readable medium carrying one or more sequences of instructions, wherein execution of the one or more sequences of instructions by one or more processors causes the one or more processors to perform the steps of claim 4 .
12. A system comprising:
a medical imaging device;
a display device;
at least one processor; and
at least one memory including one or more sequences of instructions,
the at least one memory and the one or more sequences of instructions configured to, with the at least one processor, cause an apparatus to perform at least the steps of claim 4 .