Method, apparatus and storage medium for detecting cardio, thoracic and diaphragm borders

• US 7,965,893 B2
• Filed: 06/15/2006
• Issued: 06/21/2011
• Est. Priority Date: 06/15/2005
• Status: Expired due to Fees

• Current Assignee: Canon Kabushiki Kaisha (Canon Inc.)
• Original Assignee: Canon Kabushiki Kaisha (Canon Inc.)
• Inventors: Peng, Miao, Wu, Dawei
• Primary Examiner(s): Ahmed; Samir A
• Assistant Examiner(s): Li; Ruiping
• Application Number:
  US11/424,346
• Publication Number:
  US 20060285751A1
• Time in Patent Office: 1,832 Days
• Field of Search: 382/128, 382/176, 382/199, 382/132
• US Class Current: 382/199
• CPC Class Codes: G06T 2207/10116 X-ray image
• CPC Class Codes: G06T 2207/20132 Image cropping
• CPC Class Codes: G06T 2207/30048 Heart; Cardiac
• CPC Class Codes: G06T 2207/30061 Lung
• CPC Class Codes: G06T 7/12 Edge-based segmentation
• CPC Class Codes: G06V 10/26 Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion

Abstract

The present invention provides a method of detecting borders in a chest image. The method comprises calculating the gray level difference between neighboring pixels of the chest image to get a differential image; filtering the differential image to reduce noise and enhance the target; binarizing the filtered image; picking out the outside edges of the area with high gray level from the binarized image; and determining a final border from the edges according to the character of the border to be found. The present invention also provides an apparatus and a storage medium for implementing the above-mentioned method and a method and an apparatus for automatically measuring the Cardiothoracic Ratio of a chest image.

Claims

1. A method of detecting borders in a chest image, comprising steps of:

calculating the gray level difference between neighboring pixels of the chest image to get a differential image;

filtering the differential image to reduce noise and enhance target by at least one filter;

binarizing the filtered image;

picking out outside edges of an area with high gray level from the binarized image; and

determining a final border from the edges according to the character of a border to be found,

wherein said step of determining a final border from the edges is to determine a right heart border and a left heart border by detecting a right ribcage border and a left ribcage border, detecting a right diaphragm border and a left diaphragm border on the basis of the detected right ribcage border and the detected left ribcage border, and detecting the right heart border and the left heart border on the basis of the detected right diaphragm border and the detected left diaphragm border, to detect a right heart border so that a height difference between a right heart borderline candidate center and a right diaphragm borderline center is not less than 1/32 of the chest image height, and the horizontal difference between the heart borderline candidate center and the point that is the nearest point of the diaphragm borderline from a midline is less than ⅙

of the chest image width.

2. The method according to claim 1, further comprising a step of finding landmark lines of the chest image according to a project profile of the chest image, before said step of calculating the gray level difference.

3. The method according to claim 1, wherein said step of calculating the gray level difference between neighboring pixels of the chest image comprises:

determining that the direction for calculating the gray level difference between neighboring pixels is from the brighter side of a border to be found to the darker side.

4. The method according to claim 1, wherein said step of calculating the gray level difference between neighboring pixels of the chest image comprises steps of:

transforming the gray value of the differential image into 0˜

L gray value space by linear transformation; and

computing a histogram of the differential image to get a maximum value position MP and stretching the histogram from MP˜

L to 0˜

L.

5. The method according to claim 1, wherein said step of filtering the differential image is performed using an average filter and/or a median filter.

6. The method according to claim 5, wherein the filter is chosen according to a border'"'"'s character.

7. The method according to claim 1, wherein said step of binarizing the filtered image comprises steps of:

defining a region of interest (ROI) according a midline of the chest image;

defining a binary threshold according to the average gray value of the ROI; and

binarizing the filtered image according to the binary threshold.

8. The method according to claim 1, wherein said step of picking out the outside edges comprises the steps of:

determining on which side of a high gray level area a border to be found is located;

finding out the pixels on the edges of the determined side of the high gray level area;

setting the gray level of pixels found on the edges into highest gray level and setting the gray level of other pixels into lowest gray level; and

dilating the image by an operator to ensure connectivity of the edges.

9. The method according to claim 8, wherein the operator is determined according to the character of a border to be found.

10. The method according to claim 1, wherein said step of determining a final border comprises the steps of:

tracing the outside edges to get candidate borders; and

choosing a curve in the candidate borders that has the most proper position information to be the correct border according to the character of former found borders and the character of the border to be found.

20. A non-transitory computer-readable storage medium storing a computer program which implements a method according to any one of claims 1 to 10.

11. An apparatus for detecting borders in a chest image, comprising:

means for calculating the gray level difference between neighboring pixels of the chest image to get a differential image;

means for filtering the differential image to reduce noise and enhance target;

means for binarizing the filtered image;

means for picking out outside edges of an area with high gray level from the binarized image; and

means for determining a final border from the edges according to the character of the border to be found,

wherein said means for determining a final border from the edges is to determine a right heart border and a left heart border by detecting a right ribcage border and a left ribcage border, detecting a right diaphragm border and a left diaphragm border on the basis of the detected right ribcage border and the detected left ribcage border, and detecting the right heart border and the left heart border on the basis of the detected right diaphragm border and the detected left diaphragm border, to detect a right heart border so that a height difference between a right heart borderline candidate center and a right diaphragm borderline center is not less than 1/32 of the chest image height, and the horizontal difference between the heart borderline candidate center and the point that is the nearest point of the diaphragm borderline from a midline is less than ⅙

of the chest image width.

12. The apparatus according to claim 11, wherein said means for calculating the gray level difference between neighboring pixels of the chest image comprises:

means for determining which side of a border to be found is brighter; and

means for determining that the direction for calculating the gray level difference between neighboring pixels is from the brighter side of the border to be found to the darker side.

13. The apparatus according to claim 11, wherein said means for calculating the gray level difference between neighboring pixels of the chest image comprises:

means for transforming the gray value of the differential image into 0˜

L gray value space by linear transformation; and

means for computing a histogram of the differential image to get the maximum value position MP and stretching the histogram from MP˜

L to 0˜

L.

14. The apparatus according to claim 11, wherein the means for filtering the differential image comprises an average filter or a median filter.

15. The apparatus according to claim 14, wherein said filter is chosen according to a border'"'"'s character.

16. The apparatus according to claim 11, wherein said means for binarizing the filtered image comprises:

means for defining a region of interest (ROI) according to a midline of the chest image;

means for defining a binary threshold according to the average gray value of the ROI; and

means for binarizing the filtered image according to the binary threshold.

17. The apparatus according to claim 11, wherein said means for picking out the outside edges comprises:

means for determining on which side of a high gray level area a border to be found is located;

means for finding out the pixels on the edges of the determined side of the high gray level area;

means for setting the gray level of pixels found on the edges into highest gray level and setting the gray level of other pixels into lowest gray level; and

means for dilating the image by an operator to ensure connectivity of the edges.

18. The apparatus according to claim 17, wherein the operator is determined according to the character of the border to be found.

19. The apparatus according to claim 11, wherein said means for determining a final border comprises:

means for tracing the outside edges to get the candidate borders; and

means for choosing a curve in the candidate borders that has the most proper position information to be the correct border.