Method and apparatus for automatically locating a region of interest in a radiograph

• US 5,982,916 A
• Filed: 09/30/1996
• Issued: 11/09/1999
• Est. Priority Date: 09/30/1996
• Status: Expired due to Term

• Current Assignee: Siemens AG
• Original Assignee: Siemens Corporate Research Incorporated (Siemens AG)
• Inventors: Kuhn, Gary
• Primary Examiner(s): Johns, Andrew W.
• Application Number:
  US08/724,112
• Time in Patent Office: 1,135 Days
• Field of Search: 382/128, 382/132, 382/171, 382/172, 382/270, 382/271, 382/298, 382/299, 128/922
• US Class Current: 382/132
• CPC Class Codes: G06T 2207/10116 X-ray image
• CPC Class Codes: G06T 2207/20132 Image cropping
• CPC Class Codes: G06T 2207/30068 Mammography; Breast
• CPC Class Codes: G06T 7/0012 Biomedical image inspection
• CPC Class Codes: G06T 7/11 Region-based segmentation
• CPC Class Codes: G06T 7/155 involving morphological operators
• CPC Class Codes: G06T 7/70 Determining position or orientation of objects or cameras camera calibration G06T7/80
• CPC Class Codes: G06V 10/25 Determination of region of interest [ROI] or a volume of interest [VOI]
• CPC Class Codes: G06V 2201/03 Recognition of patterns in medical or anatomical images

Abstract

A method for automatically locating a desired region of interest (ROI) in an original digital image involving the use of histograms to determine a threshold that is used to generate a mask to detect the ROIs. Digital image intensity data is obtained and a histogram of the intensity data is formed. The threshold is determined by first determining a histogram point that represents the noise peak in the image intensity data and then setting the threshold to be the image intensity value following the noise peak that first exhibits a positive curvature. The threshold is used to create a mask from the original digital image, and contiguous areas of pixels are defined in the mask. At least one of these contiguous areas is determined to be a desired ROI.

Claims

1. A method for automatically locating a desired region of interest (ROI) in an original digital image having a plurality of pixel intensity values, comprising:

(a) obtaining digital image intensity data representative of an original digital image;

(b) forming a histogram using said image intensity data, wherein one axis of said histogram indicates image intensity value and an orthogonal axis of said histogram indicates image pixel count;

(c) determining a threshold image intensity value from said histogram as being representative of a minimum image intensity value that can possibly be from a desired ROI by;

(c1) determining a point in said histogram representative of a noise peak in said image intensity data; and

(c2) processing said histogram so as to find as said threshold intensity value that image intensity value wherein a portion of said processed histogram that follows said noise peak first exhibits a change to a positive curvature;

(d) creating a mask image from said original digital image, said mask image having image pixels at grid coordinate positions corresponding to similarly positioned pixels in said original digital image, by setting all image pixels in the mask image grid to one of at least a maximum and a minimum intensity value, depending upon a comparison of the threshold intensity value to the image intensity value of the similarly positioned pixel in the original digital image; and

(e) processing said mask image so as to define at least one contiguous area of pixels in said mask image having a common intensity value, said at least one area comprising the desired ROI.

2. The method in accordance with claim 1, wherein:

step (e) defines in said mask image a plurality of areas of contiguous pixels having a common intensity value as possible ROI'"'"'s, and includes an area reduction step for processing of said mask image so as to expand background regions that either border or are internal to all possible ROI'"'"'s, thereby reducing the chances that adjacent possible ROI'"'"'s are not separately defined.

3. The method according to claim 2, wherein step (e) includes a selecting step for selecting as the desired ROI the largest of the possible ROI'"'"'s remaining in the mask image after said area reduction step.

4. The method according to claim 2, wherein said area reduction step also includes a shape eliminator step for eliminating from said mask image those possible ROI'"'"'s having shapes known to not exist in the desired ROI.

5. The method according to claim 4, wherein said shape eliminator step comprises processing of the intensity values in the mask image on a row-by-row and column-by-column basis in order to delete as possible ROI'"'"'s those having predetermined shapes.

6. The method of claim 5, wherein said area reduction step comprises expanding the background regions in the mask image by setting the intensity value of each image pixel within a possible ROI that is adjacent to a further image pixel so as to define a boundary edge therebetween, to the intensity value not used in the area of the possible ROI.

7. The method of claim 6, wherein step (e) includes a selecting step for selecting as the desired ROI the largest of the possible ROI'"'"'s remaining in the mask image after performing said area reduction step.

8. The method of claim 7, wherein step (e) includes a dilation step for dilating edges both internal and external of the desired ROI by setting the intensity value of each image pixel in the mask image in groups of image pixels that surround an image pixel having an intensity value opposite to that used in step (d) for creating the area of each ROI, to the intensity value used in the area of each ROI.

9. The method of claim 1, wherein step (a) includes the step of applying the digital intensity values used for forming said histogram to a look-up table for compressing the image intensity data.

10. The method of claim 9, wherein said look-up table is adaptive so as to apply a predetermined compression to said image intensity values independent of the bit-depth of the image intensity values.

11. The method according to claim 1, wherein step (a) includes a data reducing step for forming a reduced image having a reduced set of image intensity data as compared to a full set of image intensity data representative of said original digital image.

12. The method according to claim 11, wherein step (a) includes the step of applying a log compression processing to said image intensity data used for forming said histogram before performing said data reducing step.

13. The method according to claim 11, wherein said reducing step is performed by adaptive sub-sampling of the full set of image intensity data representative of said original digital image so as to result in a predetermined resolution of the original digital image.

14. The method of claim 1, wherein step (c1) includes the step of producing a first smoothed histogram before determining the noise peak.

15. The method of claim 14, wherein the noise peak is determined by finding the first occurrence of a fall in pixel count value in said first smoothed histogram after a given delta.

16. The method of claim 14, wherein step (c2) comprises:

producing a further smoothed histogram;

subtracting from said further smoothed histogram a less smoothed histogram so as to create a difference histogram (DH); and

selecting as said positive curvature portion that point in the difference histogram corresponding to a first increase in pixel count after the pixel count corresponding to the noise peak found by step (c1).

17. The method of claim 16, wherein said first increase is determined adaptively, by looking for said increase in a stepwise manner, said steps having a width corresponding to a predetermined percentage of the difference between a maximum image intensity value of said histogram and the intensity value corresponding to the noise peak found by step (c1).

18. The method of claim 17, further including a bias step for shifting the point of the first increase by a predetermined amount.

19. Apparatus for automatically locating a desired region of interest (ROI) in an original digital image having a plurality of pixel intensity values, comprising:

(a) means for obtaining digital image intensity data representative of an original digital image;

(b) means for forming a histogram using said image intensity data, wherein one axis of said histogram indicates image intensity value and an orthogonal axis of said histogram indicates image pixel count;

(c) means for determining a threshold image intensity value from said histogram as being representative of a minimum image intensity value that can possibly be from a desired ROI by;

(c1) determining a point in said histogram representative of a noise peak in said image intensity data; and

(c2) processing said histogram so as to find as said threshold intensity value that image intensity value wherein a portion of said processed histogram that follows said noise peak first exhibits a change to a positive curvature;

(d) means for creating a mask image from said original digital image, said mask image having image pixels at grid coordinate positions corresponding to similarly positioned pixels in said original digital image, by setting all image pixels in the mask image grid to one of at least a maximum and a minimum intensity value, depending upon a comparison of the threshold intensity value to the image intensity value of the similarly positioned pixel in the original digital image; and

(e) means for processing said mask image so as to define at least one contiguous area of pixels in said mask image having a common intensity value, said at least one area comprising the desired ROI.

20. Apparatus in accordance with claim 19, wherein said means for processing defines in said mask image a plurality of areas of contiguous pixels having a common intensity value as possible ROI'"'"'s, and includes a further means for further processing of said mask image so as to expand background regions that either border or are internal to all possible ROI'"'"'s, thereby reducing the chances that adjacent possible ROI'"'"'s are not separately defined.

21. Apparatus in accordance with claim 20, including a selecting means for selecting as the desired ROI the largest of the possible ROI'"'"'s remaining in the mask image after said further processing of said mask image.

22. Apparatus in accordance with claim 21, wherein said further processing also includes processing for eliminating from said mask image those possible ROI'"'"'s having shapes known to not exist in the desired ROI.

23. Apparatus in accordance with claim 19, wherein said means for obtaining includes image compression means for applying the digital intensity values used for forming said histogram to a look-up table for compressing the image intensity data.

24. Apparatus in accordance with claim 23, wherein said look-up table is programmed to apply a predetermined compression to said image intensity values independent of the bit-depth of the image intensity values.

25. Apparatus in accordance with claim 19, wherein said means for obtaining includes means for forming a reduced image having a reduced set of image intensity data as compared to a full set of image intensity data representative of said original digital image.

26. Apparatus for automatically locating a desired region of interest (ROI) in an original digital image having a plurality of pixel intensity values, comprising:

(a) means for obtaining digital image intensity data representative of an original digital image;

(b) means for compressing said digital image intensity data;

(c) means for sampling the digital image intensity data of said original digital image so as to create a reduced digital image representative of said original digital image, said reduced digital image comprising a sub-set of the digital image intensity data of the original digital image;

(d) means for forming a histogram using said sub-set of image intensity data, wherein one axis of said histogram indicates increasing image intensity value and an orthogonal axis of said histogram indicates increasing image pixel count;

(e) means for determining a threshold image intensity value from said histogram as being representative of a minimum image intensity value that can possibly be from a desired ROI by;

(e1) determining a point in said histogram representative of a noise peak in said image intensity data; and

(e2) processing said histogram so as to find as said threshold intensity value that image intensity value wherein a portion of said processed histogram that follows said noise peak first exhibits a change to a positive curvature; and

(f) means for creating a mask image from said sub-set of digital image intensity data, said mask image having image pixels at grid coordinate positions corresponding to similarly positioned pixels in said original digital image that were sampled for forming said reduced digital image, by setting all image pixels in a mask image grid to one of at least a maximum and a minimum intensity value, depending upon a comparison of the threshold intensity value to the image intensity value of a correspondingly positioned pixel in the reduced digital image, thereby defining at least one contiguous area of pixels in said mask image having a common intensity value, said at least one area comprising the desired ROI.