SYSTEM FOR USING THREE-DIMENSIONAL MODELS TO ENABLE IMAGE COMPARISONS INDEPENDENT OF IMAGE SOURCE

• US 20110255746A1
• Filed: 12/10/2009
• Published: 10/20/2011
• Est. Priority Date: 12/24/2008
• Status: Active Grant

• Current Assignee: Rafael Advanced Defense Systems Limited (Government of Israel)
• Original Assignee: Rafael Advanced Defense Systems Limited (Government of Israel)
• Inventors: Shapira, Dror, Berkovich, Erez
• Granted Patent: US 8,538,078 B2
• Time in Patent Office: Days
• US Class Current: 382/103
• CPC Class Codes: G06F 16/583 using metadata automatically derived from the content
• CPC Class Codes: G06T 17/00 Three dimensional [3D] modelling, e.g. data description of 3D objects
• CPC Class Codes: G06V 20/653 by matching three-dimensional models, e.g. conformal mapping of Riemann surfaces

Abstract

A method for identifying an object based at least in part on a reference database including two-dimensional images of objects includes the following steps: (a) providing a three-dimensional model reference database containing a plurality of estimated three-dimensional models, wherein each estimated three-dimensional model is derived from a corresponding two-dimensional image from the two-dimensional reference database; (b) sampling at least one image of an object to be identified; (c) implementing at least one identification process to identify the object, the identification process employing data from the three-dimensional model reference database.

Claims

1. A method for identifying an object based at least in part on a two-dimensional reference database including two-dimensional images of a plurality of objects, the method comprising:

(a) providing a three-dimensional model reference database containing a plurality of estimated three-dimensional models, wherein each estimated three-dimensional model is derived from a corresponding two-dimensional image from the two-dimensional reference database;

(b) sampling at least one image of an object to be identified; and

(c) implementing at least one identification process to identify said object, said identification process employing data from said three-dimensional model reference database.

2. The method of claim 1, wherein providing said three-dimensional model reference database further comprises generating said three-dimensional model reference database by processing the two-dimensional images to generate said corresponding estimated three-dimensional model for each of the two-dimensional images.

3. The method of claim 1, wherein said at least one sampled image is two-dimensional and further comprising processing said at least one sampled image to generate a corresponding estimated three-dimensional model of said object.

5. The method of claim 1, wherein said at least one sampled image is a plurality of two-dimensional images of the object of interest, each of said plurality of two-dimensional images having been sampled with different angles and further comprising processing said plurality of two-dimensional images to generate a corresponding estimated three-dimensional model of said object.

6. The method of claim 1, wherein said two-dimensional images from the reference database are color images and said at least one sampled image is an infrared image.

7. The method of claim 1, wherein said at least one sampled image derives from a range sensing device, and further comprising processing said sampled image to generate a corresponding estimated three-dimensional model of said object.

8. The method of claim 1, wherein the identification process employs data from said three-dimensional model reference database to create a constructed two-dimensional image to perform a two-dimensional to two-dimensional comparison.

4. The method of claim I wherein said at least one sampled image is a single two-dimensional image and further comprising processing said single two-dimensional image in conjunction with a three-dimensional generic model of the object of interest to generate a corresponding estimated three-dimensional model of said object.

9. A method for identifying an object that is independent of the source used to derive the image of said object, the method comprising:

(a) providing a two-dimensional reference image having a given texture content;

(b) sampling at least one image of an object to be identified wherein the sampled image is not compatible with the given texture content of said two-dimensional reference image;

(c) processing said two-dimensional reference image to generate a corresponding estimated three-dimensional reference model;

(d) processing said sampled image to generate a corresponding estimated three-dimensional model of the object to be identified; and

(e) implementing at least one identification process to identify said object, said identification process employing data from said estimated three-dimensional models.

10. The method of claim 9, wherein said reference image is a color image and said sampled image is an infrared image.

11. The method of claim 9, wherein said sampling is performed using a range sensing device.

12. The method of claim 9, wherein said reference image is provided from a database.

13. A method for creating a three-dimensional facial model based at least in part on an image of part of a face, the method comprising:

(a) providing a two-dimensional image of part of a face;

(b) processing said two-dimensional image of part of a face to generate a three-dimensional image of said part of face; and

(c) processing said three-dimensional image of said part of said face using symmetry to generate a full three-dimensional facial model.

14. A system for identifying an object based at least in part on a two-dimensional reference database including two-dimensional images of a plurality of objects, the system comprising:

(a) a three-dimensional model reference database, containing a plurality of estimated three-dimensional models, wherein each estimated three-dimensional model is derived from a corresponding two-dimensional images from the two-dimensional reference database;

(b) an image-sampling device configured for sampling at least one image of a real scene including the object of interest; and

(c) a processing system, including at least one processor, operationally connected to said three-dimensional model reference database, and operationally connected to said image-sampling device, said processing system configured to implement at least one identification process to identify said object.

16. The system of claim 14, wherein said processing system is configured to process said at least one sampled image, wherein said at least one sampled image is two-dimensional, to generate a corresponding estimated three-dimensional model of said object.

17. The system of claim 14, wherein said processing system is configured to process said at least one sampled image, wherein said at least one sampled image is two-dimensional, in conjunction with a three-dimensional reference model of the object of interest to generate a corresponding estimated three-dimensional model of said object.

18. The system of claim 14, wherein said processing system is configured to process said at least one sampled image, wherein said at least one sampled image is plurality of two-dimensional images of the object of interest, each of said plurality of two-dimensional images having been sampled with different angles, and further comprising processing said plurality of two-dimensional images to generate a corresponding estimated three-dimensional model of said object.

19. The system of claim 14, wherein said processing system is configured to process said at least one sampled image, wherein said at least one sampled image derives from an infrared image, in conjunction with at least one image from said reference database, wherein said at least one image from said reference database derives from a color image.

20. The system of claim 14, wherein said processing system is configured to process said at least one sampled image, wherein said at least one sampled image derives from a range sensing device.

21. The system of claim 14, wherein said processing system is configured to process data from said three-dimensional model reference database to create a constructed two-dimensional image, and wherein said at least one sampled image is two-dimensional, and further comprising processing said constructed two-dimensional image in conjunction with said sampled image to perform a two-dimensional to two-dimensional comparison.

15. A system for generating a three-dimensional model reference database by processing two-dimensional images to generate a corresponding estimated three-dimensional model for each of the two-dimensional images, the system comprising:

(a) a two-dimensional reference database including two-dimensional images of a plurality of objects; and

(b) a processing system including at least one processor, operationally connected to said two-dimensional reference database, configured to;

(i) process said two-dimensional images to generate said corresponding estimated three-dimensional model for each of said objects; and

(ii) store data representative of said estimated three-dimensional models in said three-dimensional model reference database.

22. A system for identifying an object that is independent of the source used to derive the image of said object, the system comprising:

(a) a two-dimensional reference image having a given texture content;

(b) an image-sampling device configured for sampling at least one image of an object to be identified, wherein the texture content of the sampled image is not compatible with the given texture content of said two-dimensional reference image; and

(c) a processing system including at least one processor, operationally connected to said image-sampling device configured to;

(i) process said two-dimensional reference image to generate a corresponding estimated three-dimensional reference model;

(ii) process said sampled image to generate a corresponding estimated three-dimensional model of the object to be identified;

(iii) implement at least one identification process to identify said object, said identification process employing data from said estimated three-dimensional models.

23. The system of claim 22, wherein said reference image is a color image and said sampled image is an infrared image.

24. The system of claim 22, wherein said processing system is configured to process said at least one sampled image, wherein said at least one sampled image derives from a range sensing device.

25. The system of claim 22, wherein said reference image is provided from a database.

26. A system for creating a three-dimensional facial model based at least in part on an image of part of a face, the system comprising:

(a) a two-dimensional image of part of a face;

(b) a processing system including at least one processor, configured to;

(i) process said two-dimensional image of part of a face to generate a three-dimensional image of said part of face; and

(ii) process said three-dimensional image of said part of face using symmetry to generate a full three-dimensional facial model.