Granted Patent B2

US 8,189,886 · App. 12/190,616 · Granted May 29, 2012

Abstract

Claims (32)

1. A method for estimating a location of an anatomical structure in a diagnostic image of a patient, comprising:

detecting a first benchmark feature within an x-ray image in a digital format;

detecting a second benchmark feature within the x-ray image;

applying, by a processor, at least one template to the x-ray image to determine a curvature of the second benchmark feature;

locating an intersection between a first line that extends along a length of the first benchmark feature and a second line that extends through a central point related to the curvature of the second benchmark feature and that intersects with the first line at an angle that is within a predetermined range of angles; and

identifying the location of the anatomical structure relative to the intersection.

2. The method of claim 1 wherein the first benchmark feature is a trachea.

3. The method of claim 1 wherein the first benchmark feature is a tube inserted into a trachea.

4. The method of claim 1 wherein the first benchmark feature is a spine of the patient.

5. The method of claim 1 wherein the first benchmark feature is substantially linear.

6. The method of claim 1 further comprising displaying either or both of the first line and the second line on a display.

7. The method of claim 1 wherein the second benchmark feature is an aortic arch of the patient.

8. The method of claim 7 wherein detecting the aortic arch comprises:

identifying an inner edge of a left lung within the x-ray image; and

processing the inner edge using the at least one template.

9. The method of claim 7 wherein detecting the second benchmark feature further comprises segmenting lungs within the x-ray image.

10. The method of claim 7 wherein identifying an inner edge of a left lung within the x-ray image comprises applying a Canny edge detection algorithm.

11. The method of claim 7 further comprising displaying the aortic arch on a display.

12. A method for detecting a curved feature within an image comprising:

analyzing, by a processor, the image to isolate a feature in the image;

identifying an edge of the feature within the image;

identifying successive subsections of the identified edge by subsectioning the identified edge according to a predetermined subsection dimension;

fitting one or more arcs to each subsection of the identified edge, wherein each of the one or more fitted arcs has a center of radius that lies to the same side of the identified edge and an arc radius that is within a predetermined range of radius values;

clustering two or more fitted arcs according to their proximity relative to each other;

forming a fitted edge by combining portions of the two or more clustered fitted arcs for the identified edge; and

fitting at least one additional arc to the fitted edge.

13. The method of claim 12 wherein the feature corresponds to a lung and wherein the identified edge corresponds to an inner edge of the lung.

14. A method for estimating a location of a carina in a chest x-ray image of a patient, comprising:

obtaining x-ray data in a digital format;

analyzing, by a processor, the x-ray data to identify a trachea and lungs in the x-ray data and to determine a curvature of an aortic arch from the lungs identified in the x-ray data, wherein a mask is utilized to identify lungs; and

locating an intersection between a first line that traces a length of the trachea and a second line that extends toward the first line from a center point related to the radius of curvature of the aortic arch and intersects the first line at an angle that is between 54 and 36 degrees.

15. The method of claim 14 wherein the first line is formed using a path of a spine identified in the x-ray data.

Assignments (8)