Techniques for enabling or establishing the use of face recognition algorithms

Patent metadata

Field	Value
Publication number	US-9690979-B2
Application number	US-201414153867-A
Country	US
Kind code	B2
Filing date	Jan 13, 2014
Priority date	Mar 12, 2006
Publication date	Jun 27, 2017
Grant date	Jun 27, 2017

Abstract

Embodiments described herein facilitate or enhance the implementation of image recognition processes which can perform recognition on images to identify objects and/or faces by class or by people.

First claim

What is claimed is:

1. A computer implemented method comprising:
identifying an object in an image;
obtaining, from a database of predetermined object models, an object model for the object identified in the image, the object model comprising a plurality of three dimensional points from which a three dimensional surface may be computed;
determining, using the image and the object model, an orientation and one or more occluded features of the object in the image;
augmenting the object model to compensate for the one or more occluded features of the object in the image, wherein augmenting the object model comprises:
augmenting an intensity value associated with at least one of the plurality of three dimensional points; and
rotating the object model; and

providing the augmented object model as output.

2. The method of claim 1 , wherein identifying an object in an image comprises:
identifying one or more image markers included in the image, each image marker comprising one or more predetermined image features;
determining that at least some of the one or more image markers match predefined markers of the object model; and
identifying the object in the image as an object represented by the object model.

3. The method of claim 2 , wherein determining an orientation of the object in the image comprises:
determining, for each of the one or more image markers, a relationship between the image marker and a matching predefined marker matched by the image marker; and
using each relationship to determine the orientation of the object in the image.

4. The method of claim 1 , wherein augmenting an intensity value associated with at least one of the plurality of three dimensional points comprises:
using one or more intensity values of points in the image to augment intensity values of respective three dimensional points.

5. The method of claim 1 , further comprising:
prior to augmenting the object model, orienting the object model to pose the object model at an orientation that matches the orientation of the object in the image.

6. The method of claim 1 , wherein rotating the object model comprises rotating the object model to a predefined orientation.

7. The method of claim 1 , wherein providing the augmented object model as output comprises providing the augmented object model as output to an object recognition system.

8. A system comprising:
one or more data processing apparatus; and
a data storage device storing instructions that, when executed by the one or more data processing apparatus, cause the one or more data processing apparatus to perform operations comprising:
identifying an object in an image;
obtaining, from a database of predetermined object models, an object model for the object identified in the image, the object model comprising a plurality of three dimensional points from which a three dimensional surface may be computed;
determining, using the image and the object model, an orientation and one or more occluded features of the object in the image;
augmenting the object model to compensate for the one or more occluded features of the object in the image, wherein augmenting the object model comprises:
augmenting an intensity value associated with at least one of the plurality of three dimensional points; and
rotating the object model; and

providing the augmented object model as output.

9. The system of claim 8 , wherein identifying an object in an image comprises:
identifying one or more image markers included in the image, each image marker comprising one or more predetermined image features;
determining that at least some of the one or more image markers match predefined markers of the object model; and
identifying the object in the image as an object represented by the object model.

10. The system of claim 9 , wherein determining an orientation of the object in the image comprises:
determining, for each of the one or more image markers, a relationship between the image marker and a matching predefined marker matched by the image marker; and
using each relationship to determine the orientation of the object in the image.

11. The system of claim 8 , wherein augmenting an intensity value associated with at least one of the plurality of three dimensional points comprises:
using one or more intensity values of points in the image to augment intensity values of respective three dimensional points.

12. The system of claim 8 , wherein the operations further comprise:
prior to augmenting the object model, orienting the object model to pose the object model at an orientation that matches the orientation of the object in the image.

13. The system of claim 8 , wherein rotating the object model comprises rotating the object model to a predefined orientation.

14. The system of claim 8 , wherein providing the augmented object model as output comprises providing the augmented object model as output to an object recognition system.

15. A non-transitory computer readable medium storing instructions that, when executed by one or more data processing apparatus, cause the one or more data processing apparatus to perform operations comprising:
identifying an object in an image;
obtaining, from a database of predetermined object models, an object model for the object identified in the image, the object model comprising a plurality of three dimensional points from which a three dimensional surface may be computed;
determining, using the image and the object model, an orientation and one or more occluded features of the object in the image;
augmenting the object model to compensate for the one or more occluded features of the object in the image, wherein augmenting the object model comprises:
augmenting an intensity value associated with at least one of the plurality of three dimensional points; and
rotating the object model; and

providing the augmented object model as output.

16. The non-transitory computer readable medium of claim 15 , wherein identifying an object in an image comprises:
identifying one or more image markers included in the image, each image marker comprising one or more predetermined image features;
determining that at least some of the one or more image markers match predefined markers of the object model; and
identifying the object in the image as an object represented by the object model.

17. The non-transitory computer readable medium of claim 16 , wherein determining an orientation of the object in the image comprises:
determining, for each of the one or more image markers, a relationship between the image marker and a matching predefined marker matched by the image marker; and
using each relationship to determine the orientation of the object in the image.

18. The non-transitory computer readable medium of claim 15 , wherein augmenting an intensity value associated with at least one of the plurality of three dimensional points comprises:
using one or more intensity values of points in the image to augment intensity values of respective three dimensional points.

19. The non-transitory computer readable medium of claim 15 , wherein the operations further comprise:
prior to augmenting the object model, orienting the object model to pose the object model at an orientation that matches the orientation of the object in the image.

20. The method of claim 1 , wherein determining the one or more occluded features of the object in the image comprises identifying one or more surface-ray intersections in the image.

Assignees

• Google Inc

Inventors

• Gokturk Salih Burak
• Anguelov Dragomir
• Torresani Lorenzo
• Vanhoucke Vincent O
• Shah Munjal
• Vu Diem Thanh
• Lee Kuang-Chih

Classifications

• G06V10/772

  Determining representative reference patterns, e.g. averaging or distorting patterns; Generating dictionaries · CPC title

• G06V40/171

  Local features and components; Facial parts (eye characteristics G06V40/18); Occluding parts, e.g. glasses; Geometrical relationships · CPC title

• G06T7/75 Primary

  involving models · CPC title

• G06F18/2113

  by ranking or filtering the set of features, e.g. using a measure of variance or of feature cross-correlation · CPC title

• G06F18/28

  Determining representative reference patterns, e.g. by averaging or distorting; Generating dictionaries · CPC title