Irradiation field recognition

Patent metadata

Field	Value
Publication number	US-9443160-B2
Application number	US-201213982175-A
Country	US
Kind code	B2
Filing date	Feb 27, 2012
Priority date	Mar 15, 2011
Publication date	Sep 13, 2016
Grant date	Sep 13, 2016

Abstract

A method to extract irradiation field areas in an X-ray image represented by a digital signal representation comprising the steps of segmenting the image in multiple regions of pixels which have similar local image characteristics, fitting line segments to the boundaries of these regions whereby said line segments correspond with candidate irradiation field boundaries and constitute a segmentation map, classifying regions in said segmentation map into at least two classes, one class being irradiation field and the other class being collimated region on the basis of at least one of local, regional and global image characteristics.

First claim

The invention claimed is:

1. A method of extracting irradiation field areas in an X-ray image represented by a digital signal representation, the method comprising the steps of:
first, segmenting the X-ray image into regions of pixels that have similar local image characteristics by image clustering to merge regions which have similar local image characteristics;
second, and after the first step, fitting line segments to boundaries of the merged regions so that the line segments correspond with candidate irradiation field boundaries and define a segmentation map; and
third, and after the second step, classifying regions in the segmentation map into a first class defining an irradiation field region and a second class defining a collimated region based on at least one of local, regional, and global image characteristics; wherein
a Hough transform is applied to the boundaries of the merged regions to fit the line segments corresponding with the candidate irradiation field boundaries, and a best-fit bounding box of the candidate irradiation field boundaries is determined based on a maximum line segment length in a main direction of the X-ray image and a maximum line segment length in a direction of the X-ray image perpendicular to the main direction.

2. The method according to claim 1 , wherein the step of segmenting the X-ray image into regions is based on multi-scale watershed segmentation.

3. The method according to claim 1 , wherein the image clustering includes hierarchical clustering with a measure of similarity based on at least one of a median or an average greyscale pixel value of a segmented region, a standard deviation of greyscale pixel values within a segmented region, and a position of a segmented region in the X-ray image.

4. The method according to claim 1 , wherein the Hough transform is normalized and corrected so that Hough space values of the boundaries of the candidate irradiation fields in the X-ray image have an approximate value of 1.0.

5. The method according to claim 1 , wherein only line segments are preserved that have a significant overlap with the boundaries of the merged regions.

6. The method according to claim 1 , wherein the regional image characteristics are computed out of histograms of local standard deviation of the segmented regions.

7. The method according to claim 6 , wherein one of the regional image characteristics computed out of the histograms of the local standard deviation is a distance measurement between the histograms and a histogram of local standard deviation of the total X-ray image.

8. The method according to claim 6 , wherein one of the regional image characteristics computed out of the histograms of the local standard deviation is a distance measurement between the histograms and a reference histogram of local standard deviation of only brightest regions in the X-ray image.

9. The method according to claim 6 , wherein the histograms of the local standard deviation are normalized to create normalized histograms of local standard deviation; and
one of the regional image characteristics computed out of the normalized histograms of local standard deviation is a cumulative sum of the histogram below a specified histogram abscissa.

10. The method according to claim 1 , wherein the regional image characteristic is an amount of strong edges in different segmented regions.

11. The method according to claim 1 , wherein the regional image characteristic is an average greyscale pixel difference between a region of interest and its surroundings in a neighbourhood of the boundaries of the region of interest.

12. The method according to claim 1 , wherein the step of classifying the regions into a first class and a second class defines a binary classification and is performed using a perceptron.

13. A non-transitory computer-readable recording medium comprising computer executable program code adapted to carry out the method of claim 1 when the computer executable program code is run on a computer.

Assignees

• Bertens Tom
• Agfa Healthcare Nv

Inventors

• Bertens Tom

Classifications

• G06T7/11 Primary

  Region-based segmentation · CPC title

• G06T2207/20152

  Watershed segmentation · CPC title

• G06T2207/20061

  Hough transform · CPC title

• G06T2207/10116

  X-ray image · CPC title

• G06T7/0081

  Physics · mapped topic