Automatic occlusion region identification using radiation imaging modality

Patent metadata

Field	Value
Publication number	US-9633428-B2
Application number	US-201214408283-A
Country	US
Kind code	B2
Filing date	Jun 29, 2012
Priority date	Jun 29, 2012
Publication date	Apr 25, 2017
Grant date	Apr 25, 2017

Abstract

Among other things, one or more systems and/or techniques for identifying an occlusion region in an image representative of an object subjected to examination is provided for herein. Such systems and/or techniques may find particular application in the context of object recognition analysis. An image is generated of the object and an orientation of the object is determined from the image. Based upon the determined orientation of the object relative to the direction the object is translated during examination, one or more parameters utilized for segmenting a second image of the object, identifying features in the image, and/or determining if the image comprises an occlusion region may be adjusted. In this way, the parameters utilized may be a function of the determined orientation of the object, which may mitigate false positives of detected occlusion regions.

First claim

What is claimed is:

1. An imaging apparatus for automatic occlusion detection, comprising:
a processing unit; and
memory comprising computer executable instructions that when executed by the processing unit perform operations, the operations comprising:
generating volumetric data representative of an object, the object resting on a surface during an examination from which the volumetric data is generated;
projecting the volumetric data onto an object plane parallel to the surface to generate a first two-dimensional image representative of the object;
determining an orientation of the object from the first two-dimensional image;
segmenting a second two-dimensional image representative of the object to generate a segmented image based upon one or more segmentation parameters, at least one of the one or more segmentation parameters dynamically adjusted as a function of the orientation of the object;
identifying, from the segmented image, a potential occlusion region in the second two-dimensional image; and
identifying one or more features of the potential occlusion region based upon one or more feature extraction parameters, at least one of the one or more feature extraction parameters dynamically adjusted as a function of the orientation of the object.

2. The imaging apparatus of claim 1 , the first two-dimensional image different than the second two-dimensional image.

3. The imaging apparatus of claim 1 , the second two-dimensional image depicting an object plane substantially perpendicular to the surface.

4. The imaging apparatus of claim 1 , the operations comprising:
identifying an occluded sub-object represented in the potential occlusion region by comparing the one or more features of the potential occlusion region to one or more discrimination parameters, at least one of the one or more discrimination parameters dynamically adjusted as a function of the orientation of the object.

5. The imaging apparatus of claim 1 , the determining comprising:
determining the orientation of the object by performing an Eigen analysis on the first two-dimensional image.

6. The imaging apparatus of claim 1 , the determining comprising:
defining the orientation of the object in terms of Eigen vectors and in terms of a direction of translation of the object during an examination via the imaging apparatus.

7. The imaging apparatus of claim 1 , the operations comprising:
emitting radiation toward the object; and
detecting radiation that has traversed the object.

8. The imaging apparatus of claim 1 , the operations comprising identifying a region of interest within the volumetric data, wherein the projecting the volumetric data comprises projecting merely a portion of the volumetric data representing the region of interest.

9. A method for segmenting an image to identify a potential occlusion region in the image, comprising:
generating volumetric data representative of an object, the object resting on a surface during an examination from which the volumetric data is generated;
projecting the volumetric data onto an object plane parallel to the surface to generate a first two-dimensional image representative of the object;
performing Eigen analysis on the first two-dimensional image to determine an orientation of the object;
segmenting, using one or more segmentation parameters, a second two-dimensional image representative of the object to generate a segmented image, indicative of a potential occlusion region, at least one of the one or more segmentation parameters dynamically adjusted as a function of the orientation of the object; and
identifying one or more features of the potential occlusion region based upon one or more feature extraction parameters, at least one of the one or more feature extraction parameters dynamically adjusted as a function of the orientation of the object, at least one of the generating, the projecting, the performing, the segmenting, the identifying a potential occlusion region, or the identifying one or more features of the potential occlusion region performed at least in part via a processing unit.

10. The method of claim 9 , the first two-dimensional image different than the second two-dimensional image.

11. The method of claim 9 , dynamically adjusting a segmentation parameter comprising at least one of:
dynamically adjusting a type of property utilized for segmenting the second two-dimensional image, or
dynamically adjusting a value of a property utilized for segmenting the second two-dimensional image.

12. The method of claim 9 , comprising identifying an occluded sub-object represented in the potential occlusion region by comparing the one or more features of the potential occlusion region to one or more discrimination parameters, at least one of the one or more discrimination parameters dynamically adjusted as a function of the orientation of the object.

13. The method of claim 12 , dynamically adjusting a discrimination parameter comprising at least one of:
dynamically adjusting a type of property utilized for identifying the occluded sub-object, or
dynamically adjusting a value of a property utilized for identifying the occluded sub-object.

14. The method of claim 12 , comprising:
identifying an occlusion region of the segmented image by comparing the comparing the one or more features of the potential occlusion region to one or more discrimination parameters; and
issuing an alert when an occlusion region is identified.

15. The method of claim 9 , comprising:
performing automatic threat detection on image data representative of the object when no potential occlusion region is identified in the segmented image.

16. The method of claim 9 , comprising:
performing a computed tomography (CT) examination of the object to generate the volumetric data.

17. The method of claim 9 , the object comprising baggage under examination to identify potential security threats.

18. A non-transitory computer readable medium comprising processor-executable instructions that when executed via a processing unit perform a method for automatic occlusion detection, the method comprising:
generating volumetric data representative of an object, the object resting on a surface during an examination from which the volumetric data is generated;
projecting the volumetric data onto an object plane parallel to the surface to generate a first two-dimensional image representative of the object;
determining an orientation of the object from the first two-dimensional image;
segmenting, using one or more segmentation parameters, a second two-dimensional image representative of the object to generate a segmented image, indicative of a potential occlusion region, at least one of the one or more segmentation parameters dynamically adjusted as a function of the orientation of the object; and
identifying one or more features a potential occlusion region based upon one or more feature extraction parameters, at least one of the one or more feature extraction parameters dynamically adjusted as a function of the orientation of the object.

19. The non-transitory computer readable medium of claim 18 , the method comprising:
identifying an occluded sub-object represented in the potential occlusion region by comparing the one or more features of the potential occlusion region to one or more discrimination parameters, at least one of the one or more discrimination parameters dynamically adjusted as a function of the orientation of the object.

20. The non-transitory computer readable medium of claim 18 ,

Assignees

• Simanovsky Sergey
• Naidu Ram
• Analogic Corp

Inventors

• Simanovsky Sergey
• Naidu Ram

Classifications

• G06T7/136

  involving thresholding · CPC title

• G06T2207/30112

  Baggage; Luggage; Suitcase · CPC title

• G01N23/046

  using tomography, e.g. computed tomography [CT] · CPC title

• G06T7/0004 Primary

  Industrial image inspection · CPC title

• G06T2207/10081

  Computed x-ray tomography [CT] · CPC title