Model-based segmentation of an anatomical structure

Patent metadata

Field	Value
Publication number	US-10282846-B2
Application number	US-201415039899-A
Country	US
Kind code	B2
Filing date	Dec 2, 2014
Priority date	Dec 10, 2013
Publication date	May 7, 2019
Grant date	May 7, 2019

Abstract

A method is provided for generating a deformable model ( 300 ) for segmenting an anatomical structure in a medical image. The anatomical structure comprises a wall. The deformable model ( 300 ) is generated such that it comprises, in addition to two surface meshes ( 320, 360 ), an intermediate layer mesh ( 340 ) for being applied in-between a first surface layer of the wall and a second surface layer of the wall. In generating the intermediate layer mesh ( 340 ), the mesh topology of at least part ( 400 ) of the intermediate layer mesh is matched to the mesh topology of one of the surface meshes ( 320, 360 ), thereby establishing matching mesh topologies. The deformable model ( 300 ), as generated, better matches the composition of such walls, thereby providing a more accurate segmentation.

First claim

The invention claimed is:

1. A method for generating a deformable model for segmenting an anatomical structure in a medical image the anatomical structure comprising a wall having a first surface layer and a second surface layer, the method comprising:
providing a first surface mesh for being applied to the first surface layer of the wall during a model-based segmentation;
providing a second surface mesh for being applied to the second surface layer of the wall during the model-based segmentation;
generating an intermediate layer mesh for being applied in-between the first surface layer and the second surface layer during the model-based segmentation, said generating comprising matching a mesh topology of at least part of the intermediate layer mesh to the mesh topology of an anatomically corresponding part of a selected one of both surface meshes, thereby establishing matching mesh topologies; said generating comprising duplicating the part of the selected surface mesh to obtain the part of the intermediate layer mesh;
generating model data defining the first surface mesh, the second surface mesh and the intermediate layer mesh, said generating comprising defining correspondences between original nodes of the selected surface mesh and duplicated nodes of the intermediate layer mesh, thereby establishing a plurality of node pairs;
wherein the generating of the model data further comprises minimizing an energy function penalizing undesired mesh states;
wherein the minimization of the energy function is determined by a) a deviation of edge lengths between the surface mesh and the intermediate layer mesh, orb) a spatial deviation of the duplicated nodes from the original nodes, or c) a growth of intermediate layer mesh edges.

2. The method according to claim 1 , further comprising generating a volume mesh between the intermediate layer mesh and the selected surface mesh based on the correspondences between the original nodes and the duplicated nodes of the node pairs.

3. The method according to claim 1 , further comprising generating a number of intermediate layer meshes for being applied in-between the first surface layer and the second surface layer of the wall, the number being larger than one.

4. The method according to claim 3 , wherein the number is established based on at least one of the group of: an input from a user, an anatomy of the wall, and a spatial resolution of the medical image.

5. The method according to claim 1 , wherein the intermediate layer mesh is generated for being applied to one of the group of: an intermediate anatomical layer of the wall, and an artificially defined layer.

6. The method according to claim 1 , further comprising:
providing one of: the first surface mesh and the second surface mesh; and
generating the remaining one of: the first surface mesh and the second surface mesh, said generating comprising establishing matching mesh topologies for anatomically corresponding parts of both surface meshes.

7. A non-transitory computer-readable medium comprising a computer program product comprising instructions for causing a processor system to perform the method according to claim 1 .

8. A non-transitory computer readable medium comprising instructions for causing a processor system to perform operations defining a deformable model, the deformable model being arranged for segmenting an anatomical structure in a medical image, the anatomical structure comprising a wall having a first surface layer and a second surface layer, the deformable model comprising:
a first surface mesh for being applied to the first surface layer of the wall during a model-based segmentation;
a second surface mesh for being applied to the second surface layer of the wall during the model-based segmentation; and
an intermediate layer mesh for being applied in-between the first surface layer and the second surface layer during the model-based segmentation, wherein at least part of the intermediate layer mesh comprises a mesh topology which matches the mesh topology of an anatomically corresponding part of a selected one of both surface meshes, thereby establishing matching mesh topologies, wherein the intermediate layer mesh is generated by duplicating the part of the selected surface mesh to obtain the part of the intermediate layer mesh;
wherein the deformable model is generated by defining correspondences between original nodes of the selected surface mesh and duplicated nodes of the intermediate layer mesh, thereby establishing a plurality of node pairs,
wherein the deformable model is generated by minimizing an energy function penalizing undesired mesh states, wherein the minimization of the energy function is determined by either: a) a deviation of edge lengths between the surface mesh and the intermediate layer mesh, orb) a spatial deviation of the duplicated nodes from the original nodes, or c) a growth of intermediate layer mesh edges, or any combination thereof.

9. A system for applying a deformable model to an anatomical structure in a medical image, the anatomical structure comprising a wall having a first surface layer and a second surface layer, the system comprising:
a non-transitory computer readable medium according to claim 8 ; and
a processing subsystem configured for applying the deformable model to the anatomical structure in the medical image.

10. The system according to claim 9 , wherein the processing subsystem is configured for applying the deformable model to the anatomical structure by:
applying the first surface mesh to the first surface layer of the anatomical structure and the second surface mesh to the second surface layer of the anatomical structure based on an optimization comprising an external energy term, the external energy term being defined as a function of image data of the medical image; and
applying the intermediate layer mesh in-between the first surface layer and the second surface layer of the wall based on one of the group of: use of an internal energy term in the optimization, and data being indicative of a relative position of the intermediate layer mesh with respect to one of said applied surface meshes.

11. The method according to claim 1 , wherein the generating of the model data further comprises defining distances between the original nodes and the duplicated nodes of the node pairs.

12. The method according to claim 1 , wherein the generating of the model data further comprising:
providing a first set of distances for a first subset of the node pairs;
obtaining a second set of distances for a second subset of the node pairs by interpolating the first set of distances.

Assignees

• Koninklijke Philips Nv

Inventors

• Groth Alexandra
• Nickisch Hannes
• Weber Frank Michael
• Weese Juergen
• Barschdorf Hans

Classifications

• G06T2207/30048

  Heart; Cardiac · CPC title

• G06T7/149 Primary

  involving deformable models, e.g. active contour models · CPC title

• G06T7/10

  Segmentation; Edge detection (motion-based segmentation G06T7/215) · CPC title

• A61B5/05

  Detecting, measuring or recording for diagnosis by means of electric currents or magnetic fields; Measuring using microwaves or radio waves (measuring movement of the entire body or parts thereof A61B5/11; detecting, measuring or recording bioelectric or biomagnetic signals of the body or parts thereof A61B5/24) · CPC title

• G06T2207/10072

  Tomographic images · CPC title