Systems and methods for volumetric segmentation of structures in planar medical images

1. A method for volumetric segmentation of a structure in a plurality of planar medical images, the method comprising:

receiving, at an electronic processor, the plurality of planar medical images, wherein the plurality of planar medical images form a three dimensional (3D) volume including the structure;

displaying, on a display, a first planar medical image from the plurality of planar medical images;

detecting, with a user interface, a user selection at a first location in the first planar medical image;

detecting, with the user interface, a user deselection at a second location in the first planar medical image, wherein the second location is different from the first location;

prior to detecting the user deselection,

receiving, with the user interface, a third location in the first planar medical image, wherein the third location is different from the first location and the second location;

determining, with the electronic processor, a first two dimensional (2D) contour of the structure using a first single straight line segment between the third location and the first location;

re-displaying, on the display, the first planar medical image with visual indicators of the first single straight line segment and the first 2D contour of the structure;

receiving, with the user interface, the second location after receiving the third location;

determining, with the electronic processor, a second 2D contour of the structure using a second single straight line segment between the second location and the first location; and

re-displaying, on the display, the first planar medical image with visual indicators of the second single straight line segment and the second 2D contour of the structure; and

after detecting the user deselection,

determining, with the electronic processor, an inclusion region of the 3D volume using the second single straight line segment, wherein the inclusion region consists of a portion of the structure,

determining, with the electronic processor, a containment region of the 3D volume using the second single straight line segment, wherein the containment region includes the structure,

determining, with the electronic processor, a 3D contour of the structure using the inclusion region and the containment region;

determining, with the electronic processor, a long axis of the structure using the 3D contour of the structure; and

outputting, with the electronic processor, a dimension of the long axis of the structure.

2. The method of claim 1 , wherein the dimension is a first dimension, wherein the method further comprising

determining, with the electronic processor, a short axis of the structure using the 3D contour of the structure; and

outputting, with the electronic processor, a second dimension of the short axis of the structure.

3. The method of claim 1 , further comprising:

determining, with the electronic processor, a background region of the 3D volume using the second single straight line segment, wherein the background region excludes the structure, and

determining, with the electronic processor, the 3D contour of the structure further using the background region.

4. The method of claim 3 , further comprising

classifying, with the electronic processor, a first set of voxels located within the inclusion region as belonging to a foreground class;

classifying, with the electronic processor, a second set of voxels located within the background region as belonging to a background class;

statistically sampling, with the electronic processor, the first set of voxels and the second set of voxels to classify a third set of voxels located within the containment region and outside the inclusion region as belonging to either the foreground class or the background class;

determining, with the electronic processor, a border in the 3D volume between the third set of voxels belonging to the foreground class and the third set of voxels belonging to the background class; and

determining, with the electronic processor, the 3D contour of the structure based on the border in the 3D volume.

5. The method of claim 4 , further comprising

determining, with the electronic processor, an avoidance region of the 3D volume, wherein the avoidance region does not include the structure;

classifying, with the electronic processor, a fourth set of voxels within the avoidance region as belonging to the background class; and

classifying, with the electronic processor, the third set of voxels as belonging to either the foreground class or the background class without statistically sampling the fourth set of voxels.

6. The method of claim 3 , wherein determining, with the electronic processor, the containment region using the second single straight line segment including

determining, with the electronic processor, three 2D contours of the structure using the second single straight line segment, wherein each of the three 2D contours of the structure is located in a different plane of the 3D volume, and

determining, with the electronic processor, the containment region as an ellipsoid in the 3D volume that completely encompasses the three 2D contours of the structure.

7. The method of claim 6 , wherein determining, with the electronic processor, the three 2D contours of the structure using the second single straight line segment including

statistically sampling, with the electronic processor, along the second single straight line segment to determine a probability distribution for image brightness, and

determining, with the electronic processor, the three 2D contours of the structure based on the probability distribution for image brightness.

8. The method of claim 6 , wherein determining, with the electronic processor, the three 2D contours of the structure using the second single straight line segment including

determining, with the electronic processor, boundary profiles based on the end points of the second single straight line segment, and

sampling, with the electronic processor, the boundary profiles to determine the three 2D contours of the structure.

9. A system for determining volumetric segmentation of a structure in a plurality of planar medical images, the system comprising:

a display;

a user interface; and

an electronic processor configured to

receive the plurality of planar medical images, wherein the plurality of planar medical images form a three dimensional (3D) volume including the structure and a plurality of voxels,

display, on the display, a first planar medical image from the plurality of planar medical images,

detect, with the user interface, a user selection at a first location in the first planar medical image,

detect, with the user interface, a user deselection at a second location in the first planar medical image, wherein the second location is different from the first location,

prior to detecting the user deselection,

receive, with the user interface, a third location in the first planar medical image, wherein the third location is different from the first location and the second location,

determine a first two dimensional (2D) contour of the structure using a first single straight line segment between the third location and the first location,

re-display, on the display, the first planar medical image with visual indicators of the first single straight line segment and the first 2D contour of the structure,

receive, with the user interface, the second location after receiving the third location,

determine a second 2D contour of the structure using a second single straight line segment between the second location and the first location, and

re-display, on the display, the first planar medical image with visual indicators of the second single straight line segment and the second 2D contour of the structure, and

after detecting the user deselection,

classify the voxels as belonging to either a foreground class or a background class using the second single straight line segment,

determine an inclusion region of the 3D volume using the second single straight line segment, wherein the inclusion region consists of a portion of the structure,

determine a containment region of the 3D volume using the second single straight line segment, wherein the containment region includes the structure,

determine a 3D contour of the structure based on a border in the 3D volume between the voxels belonging to the foreground class and the voxels belonging to the background class using the inclusion region and the containment region;

determine a long axis of the structure using the 3D contour of the structure, and

output a dimension of the long axis of the structure.

10. The system of claim 9 , wherein the dimension is a first dimension, wherein the electronic processor is further configured to

determine a short axis of the structure using the 3D contour of the structure, and

output a second dimension of the short axis of the structure.

11. The system of claim 9 , wherein, to classify the voxels as belonging to either the foreground class or the background class using the second single straight line segment, the electronic processor is further configured to

determine a background region of the 3D volume using the second single straight line segment, wherein the background region excludes the structure, and

classify the voxels located within the containment region as belonging to either the foreground class or the background class further using the background region.

12. The system of claim 11 , wherein the electronic processor is further configured to

classify the voxels located within the inclusion region as belonging to the foreground class,

classify the voxels located within the background region as belonging to the background class, and

statistically sample the voxels locating within the inclusion region and the voxels located within the background region to classify the voxels located within the containment region and outside the inclusion region as belonging to either the foreground class or the background class.

13. The system of claim 11 , wherein the electronic processor is further configured to

determine an avoidance region of the 3D volume, wherein the avoidance region does not include the structure,

classify the voxels located within the avoidance region as belonging to the background class, and

classify the voxels located within the containment region and outside the inclusion region as belonging to either the foreground class or the background class without statistically sampling the voxels located within the avoidance region.

14. The system of claim 11 , wherein the electronic processor is further configured to

receive, with the user interface, a user input indicating a third single straight line segment in the first planar medical image,

adjust the second 2D contour of the structure using the third single straight line segment and the second single straight line segment,

determine the inclusion region of the 3D volume using the third single straight line segment, the second single straight line segment, and the adjusted second 2D contour of the structure,

determine the containment region of the 3D volume using the third single straight line segment, the second single straight line segment, and the adjusted second 2D contour of the structure, and

determine the background region of the 3D volume using the third single straight line segment, the second single straight line segment, and the adjusted second 2D contour of the structure.

15. The system of claim 11 , wherein the electronic processor is further configured to

receive, with the user interface, a user input indicating an edit to the second 2D contour of the structure,

determine the inclusion region of the 3D volume using the second single straight line segment and the edit to the second 2D contour of the structure,

determine the containment region of the 3D volume using the second single straight line segment and the edit to the second 2D contour of the structure, and

determine the background region of the 3D volume using the second single straight line segment and the edit to the second 2D contour of the structure.

16. The system of claim 11 , wherein the electronic processor is further configured to

display, on the display, a second planar medical image from the plurality of planar medical images, wherein a first plane of the first planar medical image is different from a second plane of the second planar medical image,

receive, with the user interface, a user input indicating a third single straight line segment in the second planar medical image,

determine the inclusion region of the 3D volume using the third single straight line segment and the second single straight line segment,

determine the containment region of the 3D volume using the third single straight line segment and the second single straight line segment, and

determine the background region of the 3D volume using the third single straight line segment and the second single straight line segment.

17. The system of claim 11 , wherein the 3D contour of the structure is a first 3D contour of the structure, wherein the electronic processor is further configured to

re-display, on the display, the first planar medical image with a visual indicator of the first 3D contour of the structure,

receive, with the user interface, a user input indicating an edit to the first 3D contour of the structure,

determine a second inclusion region of the 3D volume using the second single straight line segment and the edit to the first 3D contour of the structure,

determine a second containment region of the 3D volume using the second single straight line segment and the edit to the first 3D contour of the structure,

classify the voxels located within the second containment region as belonging to either the foreground class or the background class using the second inclusion region, the second containment region, and the background region, and

determine a second 3D contour of the structure based on a second border in the 3D volume between the voxels belonging to the foreground class and the voxels belonging to the background class.

18. A non-transitory machine-readable storage medium for use in connection with a system comprising an electronic processor, a display, and a user interface, the non-transitory machine-readable storage medium comprising instructions that when executed by the electronic processor cause the electronic processor to

receive a plurality of planar medical images, wherein the plurality of planar medical images form a three dimensional (3D) volume including a structure;

display, on the display, a first planar medical image from the plurality of planar medical images;

detect, with the user interface, a user selection at a first location in the first planar medical image;

detect, with the user interface, a user deselection at a second location in the first planar medical image, wherein the second location is different from the first location;

prior to detecting the user deselection,

receive, with the user interface, a third location in the first planar medical image, wherein the third location is different from the first location and the second location;

determine a first two dimensional (2D) contour of the structure using a first single straight line segment between the third location and the first location;

re-display, on the display, the first planar medical image with visual indicators of the first single straight line segment and the first 2D contour of the structure;

receive, with the user interface, the second location after receiving the third location;

determine a second 2D contour of the structure using a second single straight line segment between the second location and the first location; and

re-display, on the display, the first planar medical image with visual indicators of the second single straight line segment and the second 2D contour of the structure; and

after detecting the user deselection,

determine an inclusion region of the 3D volume using the second single straight line segment, wherein the inclusion region consists of a portion of the structure,

determine a containment region of the 3D volume using the second single straight line segment, wherein the containment region includes the structure,

determine a 3D contour of the structure using the inclusion region and the containment region;

determine a long axis of the structure using the 3D contour of the structure; and

output a dimension of the long axis of the structure.

19. The non-transitory machine-readable storage medium of claim 18 , wherein the dimension is a first dimension, wherein the instructions further cause the electronic processor to

determine a short axis of the structure using the 3D contour of the structure; and

output a second dimension of the short axis of the structure.

20. The non-transitory machine-readable storage medium of claim 18 , wherein, to determine the 3D contour of the structure using the second single straight line segment, the instructions further cause the electronic processor to

determine a background region of the 3D volume using the second single straight line segment, wherein the background region excludes the structure, and

determine the 3D contour of the structure further using the background region.