Methods for extracting shape feature, inspection methods and apparatuses

Patent metadata

Field	Value
Publication number	US-9412019-B2
Application number	US-201314136462-A
Country	US
Kind code	B2
Filing date	Dec 20, 2013
Priority date	Dec 27, 2012
Publication date	Aug 9, 2016
Grant date	Aug 9, 2016

Abstract

Methods for extracting a shape feature of an object and security inspection methods and apparatuses. Use is made of CT's capability of obtaining a 3D structure. The shape of an object in an inspected luggage is used as a feature of a suspicious object in combination with a material property of the object. For example, a false alarm rate in detection of suspicious explosives may be reduced.

First claim

What is claimed is:

1. A method to extract a shape feature of an object in a computed tomography (CT) system, the method comprising:
acquiring slice data of an article under inspection with the CT system;
generating, from the slice data, 3-dimensional (3D) volume data of an object in the article;
calculating, based on the 3D volume data, a first depth projection image of the object in a direction perpendicular to a horizontal plane, and second, third and fourth depth projection images in three other directions, wherein a projection direction of the fourth depth projection image is orthogonal to the projection directions of the second and third depth projection images;
calculating a metric of probability that the first depth projection image might contain the horizontal plane;
calculating a metric of symmetry for each of the first, second, third, and fourth depth projection images; and
generating a shape feature parameter for the object at least based on the metric of probability and the respective metrics of symmetry of the first to fourth depth projection images.

2. The method of claim 1 , wherein the calculating the metric of probability that the first depth projection image might contain the horizontal plane comprises calculating an area ratio of a liquid surface portion to the first projection image as the metric of probability.

3. The method of claim 1 , wherein the calculating the metric of probability that the first depth projection image might contain the horizontal plane comprises calculating a shape similarity between a liquid surface portion and the first projection image as the metric of probability.

4. The method of claim 1 , wherein the calculating the metric of probability that the first depth projection image might contain the horizontal plane comprises calculating an offset that a center of gravity of a liquid surface portion deviates from a center of gravity of the first projection image as the metric of probability.

5. The method of claim 1 , wherein the calculating the metric of probability that the first depth projection image might contain the horizontal plane comprises calculating a degree that a metric of symmetry of a liquid surface portion approximates a metric of symmetry of the first projection image as the metric of probability.

6. The method of claim 1 , wherein the projection directions of the second and third depth projection images are substantially orthogonal to each other, and approximate directions of maximal and minimal projection areas of the object, respectively.

7. The method of claim 1 , further comprising calculating a similarity between each of two of the second, third and fourth depth projection images, wherein the shape feature parameter further includes the calculated similarities.

8. The method of claim 1 , further comprising calculating a duty ratio and aspect ratio for each of the second, third and fourth depth projection images, wherein the shape feature parameter further includes the duty ratio and aspect ratio.

9. The method of claim 1 , further comprising calculating a number of facets of a 3D model for the object based on the 3D volume data, and determining a complexity of the 3D model based on the number of facets and a predefined average number of facets, wherein the shape feature parameter further includes the complexity.

10. The method of claim 1 , wherein the generating, from the slice data, 3D volume data of the object comprises:
performing image segmentation on the slice data to divide them into a plurality of regions;
connecting the regions of different slice data based on relations between binary masks of the respective regions to obtain connected object data; and
performing inter-slice interpolation on the binary masks of the object to obtain the 3D volume data of the object.

11. The method of claim 1 , further comprising, prior to the calculating the first to fourth depth projection images, performing isosurface extraction and isosurface smoothing on the 3D volume data.

12. An apparatus to extract a shape feature of an object in a computed tomography (CT) system, the apparatus comprising;
a storage device configured to store slice data from inspection with the CT system;
a processing unit configured to:
acquire slice data of an article under inspection with the CT system;
generate, from the slice data, 3-dimensional (3D) volume data of an object in the article;
calculate, based on the 3D volume data, a first depth projection image of the object in a direction perpendicular to a horizontal plane, and second, third and fourth depth projection images in three other directions, wherein a projection direction of the fourth depth projection image is orthogonal to projection directions of the second and third depth projection images;
calculate a metric of probability that the first depth projection image might contain the horizontal plane;
calculate a metric of symmetry for each of the first, second, third, and fourth depth projection images; and
generate a shape feature parameter for the object at least based on the metric of probability and the respective metrics of symmetry of the first to fourth depth projection images.

13. The apparatus of claim 12 , wherein, to calculate the metric of probability that the first depth projection image might contain the horizontal plane, the processing unit is further configured to calculate an area ratio of a liquid surface portion to the first projection image as the metric of probability.

14. The apparatus of claim 12 , wherein, to calculate the metric of probability that the first depth projection image might contain the horizontal plane, the processing unit is further configured to calculate a shape similarity between a liquid surface portion and the first projection image as the metric of probability.

15. The apparatus of claim 12 , wherein, to calculate the metric of probability that the first depth projection image might contain the horizontal plane, the processing unit is further configured to calculate an offset that a center of gravity of a liquid surface portion deviates from a center of gravity of the first projection image as the metric of probability.

16. The apparatus of claim 12 , wherein, to calculate the metric of probability that the first depth projection image might contain the horizontal plane, the processing unit is further configured to calculate a degree that a metric of symmetry of a liquid surface portion approximates a metric of symmetry of the first projection image as the metric of probability.

17. The apparatus of claim 12 , wherein the projection directions of the second and third depth projection images are substantially orthogonal to each other, and approximate directions of maximal and minimal projection areas of the object, respectively.

18. The apparatus of claim 12 , wherein the processing unit is further configured to calculate a similarity between each of two of the second, third and fourth depth projection images, wherein the shape feature parameter further includes the calculated similarities.

19. The apparatus of claim 12 , wherein the processing unit is further configured to calculate a duty ratio and aspect ratio for each of the second, third and fourth depth projection images, wherein the shape feature parameter further includes the duty ratio and aspect ratio.

20. The apparatus of claim 12 , wherein the processing unit is further configured to calculate a number of facets of a 3D model for the object based on the 3D volume data, and determine a complexity of the 3D

Assignees

• Univ Tsinghua
• Nuctech Co Ltd

Inventors

• Chen Zhiqiang
• Zhang Li
• Zhao Ziran
• Huang Qingping
• Li Qiang
• Gu Jianping
• Sun Yunda

Classifications

• G06T7/596

  from three or more stereo images · CPC title

• G06T7/0004 Primary

  Industrial image inspection · CPC title

• G06F2218/12

  Classification; Matching · CPC title

• G06F2218/08

  Feature extraction · CPC title

• G06T12/30

  Image post-processing, e.g. metal artefact correction · CPC title