Threshold determination in a RANSAC algorithm

• US 10,225,473 B2
• Filed: 12/16/2016
• Issued: 03/05/2019
• Est. Priority Date: 12/17/2015
• Status: Active Grant

• Current Assignee: Stmicroelectronics France (STMicroelectronics NV)
• Original Assignee: STMicroelectronics SA (STMicroelectronics NV)
• Inventors: Alibay, Manu, Auberger, Stephane
• Primary Examiner(s): Tran, Sinh
• Assistant Examiner(s): Le, Tuan H
• Application Number:
  US15/382,013
• Publication Number:
  US 20170180644A1
• Time in Patent Office: 809 Days
• CPC Class Codes: G06T 2207/20076 Probabilistic image processing
• CPC Class Codes: G06T 7/33 using feature-based methods
• CPC Class Codes: G06T 7/35 using statistical methods
• CPC Class Codes: H04N 23/6811 based on the image signal
• CPC Class Codes: H04N 23/683 performed by a processor, e.g. controlling the readout of an image memory
• CPC Class Codes: H04N 5/145 Movement estimation for video coding H04N19/51

Abstract

A method determines a movement of an apparatus between capturing first and second images. The method includes testing model hypotheses of the movement by for example a RANSAC algorithm, operating on a set of first points in the first image and assumed corresponding second points in the second image to deliver the best model hypothesis. The testing includes, for each first point, calculating a corresponding estimated point using the tested model hypothesis, determining the back-projection error between the estimated point and the second point in the second image, and comparing each back projection error with a threshold. The testing comprises for each first point, determining a correction term based on an estimation of the depth of the first point in the first image and an estimation of the movement between the first and second images, and determining the threshold associated with the first point by using said correction term.

Claims

1. A method, comprising:

determining a movement of an imaging device between capture of successive first and second video images by said imaging device, said determining including;

testing a plurality of model hypotheses of said movement by a deterministic algorithm operating on a set of first points in the first video image and second points in the second video image so as to deliver a best model hypothesis, said testing including for each model hypothesis and for each first point, calculating a corresponding estimated point using the model hypothesis, determining a back-projection error between said estimated point and a corresponding point of the second points in the second video image, and comparing each back projection error with a threshold, wherein the testing comprises for each first point of the first video image, determining a correction term taking into account an estimation of a depth of said first point in said first video image and an estimation of the movement of the imaging device between capturing the first and the second video images, and determining the threshold associated with said first point by using said correction term.

2. The method according to claim 1, wherein said movement is an orientation modification of said imaging device between the first video image and the second video image.

3. The method according to claim 2, wherein each model hypothesis comprises Euler angles.

4. The method according to claim 1, wherein each model hypothesis is obtained from motion vectors.

5. The method according to claim 1, wherein said movement of the imaging device between capturing the first and second video images comprises a translational movement of the imaging device.

6. The method according to claim 1, wherein, for each first point, the threshold associated with the first point is a sum of a fixed threshold and said correction term for the first point.

7. A device for determining a movement of an imaging device between capturing successive first and second video images, said device comprising:

processing means for performing a testing phase of testing a plurality of model hypotheses of said movement by a deterministic algorithm operating on a set of first points in the first video image and second points in the second video image so as to deliver a best model hypothesis, said testing including, for each first point and for each model hypothesis, calculating a corresponding estimated point using the model hypothesis, determining a back-projection error between said estimated point and a corresponding point of the second points in the second video image, and comparing each back projection error with a threshold, wherein said processing means comprises calculation means for each first point of the first video image, determining a correction term taking into account a depth of said first point in said first video image and the movement of the imaging device between capturing the first and the second video images, and determining the threshold associated with said first point by using said correction term.

8. The device according to claim 7, wherein said movement is an orientation modification of said imaging device between capturing the first video image and the second video image.

9. The device according to claim 8, wherein each model hypothesis comprises Euler angles.

10. The device according to claim 7, wherein each model hypothesis is obtained from motion vectors.

11. The device according to claim 7, wherein said movement between capturing the first and second video images comprises a translational movement of the imaging device.

12. The device according to claim 7, wherein, for each first point, the threshold associated with the first point is a sum of a fixed threshold and said correction term for the first point.

13. An apparatus, comprising:

an imaging device configured to capture successive first and second video images; and

a movement detecting device configured to determine a movement of an imaging device between capturing the first and second video images, the movement detecting device including;

processing means for performing a testing phase of testing a plurality of model hypotheses of said movement by a deterministic algorithm operating on a set of first points in the first video image and second points in the second video image so as to deliver a best model hypothesis, said testing including, for each first point and for each model hypothesis, calculating a corresponding estimated point using the model hypothesis, determining a back-projection error between said estimated point and a corresponding point of the second points in the second video image, and comparing each back projection error with a threshold, wherein said processing means comprises calculation means for each first point of the first video image, determining a correction term taking into account a depth of said first point in said first video image and the movement of the imaging device between capturing the first and the second video images, and determining the threshold associated with said first point by using said correction term.

14. The apparatus according to claim 13, being a smartphone or a tablet.

15. The apparatus according to claim 13, wherein, for each first point, the threshold associated with the first point is a sum of a fixed threshold and said correction term for the first point.

16. A nontransitory computer readable medium storing instructions that, when executed by a processor, perform a method comprising:

determining a movement of an imaging device between capture of successive first and second video images by said imaging device, said determining including;

testing a plurality of model hypotheses of said movement by a deterministic algorithm operating on a set of first points in the first video image and second points in the second video image so as to deliver a best model hypothesis, said testing including for each model hypothesis and for each first point, calculating a corresponding estimated point using the model hypothesis, determining a back-projection error between said estimated point and a corresponding point of the second points in the second video image, and comparing each back projection error with a threshold, wherein the testing comprises for each first point of the first video image, determining a correction term taking into account an estimation of a depth of said first point in said first video image and an estimation of the movement of the imaging device between capturing the first and the second video images, and determining the threshold associated with said first point by using said correction term.

17. The nontransitory computer readable medium according to claim 16, wherein said movement is rotational movement of said imaging device between capturing the first video image and the second video image.

18. The nontransitory computer readable medium according to claim 16, wherein each model hypothesis is obtained from motion vectors.

19. The nontransitory computer readable medium according to claim 16, wherein said movement of the imaging device between capturing the first and second video images comprises a translational movement of the imaging device.

20. The nontransitory computer readable medium according to claim 16, wherein, for each first point, the threshold associated with the first point is a sum of a fixed threshold and said correction term for the first point.