METHOD AND APPARATUS FOR FACIAL RECOGNITION

• US 20140050372A1
• Filed: 01/18/2013
• Published: 02/20/2014
• Est. Priority Date: 08/15/2012
• Status: Active Grant

• Current Assignee: Qualcomm, Inc.
• Original Assignee: Qualcomm, Inc.
• Inventors: Qi, Yingyong, Bi, Ning, Guo, Feng
• Granted Patent: US 9,183,429 B2
• Time in Patent Office: Days
• US Class Current: 382/118
• CPC Class Codes: G06F 18/213 Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods
• CPC Class Codes: G06F 18/28 Determining representative reference patterns, e.g. by averaging or distorting; Generating dictionaries
• CPC Class Codes: G06V 10/467 Encoded features or binary features, e.g. local binary patterns [LBP]
• CPC Class Codes: G06V 10/7715 Feature extraction, e.g. by transforming the feature space, e.g. multi-dimensional scaling [MDS]; Mappings, e.g. subspace methods
• CPC Class Codes: G06V 10/772 Determining representative reference patterns, e.g. averaging or distorting patterns; Generating dictionaries
• CPC Class Codes: G06V 40/16 Human faces, e.g. facial parts, sketches or expressions
• CPC Class Codes: G06V 40/172 Classification, e.g. identification

Abstract

Apparatus and methods for facial detection are disclosed. A plurality of images of an observed face is received for identification. Based at least on two or more selected images of the plurality of images, a template of the observed face is generated. In some embodiments, the template is a subspace generated based on feature vectors of the plurality of received images. A database of identities and corresponding facial data of known persons is searched based at least on the template of the observed face and the facial data of the known persons. One or more identities of the known persons are selected based at least on the search.

Claims

1. A computer-implemented method of facial recognition, comprising:

receiving a plurality of images of an observed face for identification;

generating a template of the observed face based at least on two or more selected images of the plurality of images;

searching a database of identities and corresponding facial data of known persons, the searching being based at least on the template of the observed face and the facial data of the known persons; and

selecting one or more identities of the known persons based at least on the search.

2. The computer-implemented method of claim 1, wherein each identity of the database is associated with data that is based on one reference image.

3. The computer-implemented method of claim 1, further comprising capturing the plurality of images of the observed face with an image sensor.

4. The computer-implemented method of claim 1, wherein generating the template of the observed face includes generating a subspace of the observed face based at least on feature vectors of the two or more selected images of the plurality of images.

5. The computer-implemented method of claim 4, wherein the feature vectors are based at least on linear binary pattern histograms of the selected images.

6. The computer-implemented method of claim 4, wherein the feature vectors are based at least on a cascade of filtered image blocks.

7. The computer-implemented method of claim 4, wherein the subspace is associated with a mean and a variance of the feature vectors of the two or more selected images of the plurality of images.

8. The computer-implemented method of claim 1, wherein searching the database comprises determining an error measure between the template of the observed face and each of one or more of the corresponding facial data of the known persons.

9. The computer-implemented method of claim 8, wherein the error measure is based on a Chi-square distance.

10. The computer-implemented method of claim 8, wherein the selection of the one or more identities is based on a threshold of the error measure.

11. The computer-implemented method of claim 1, further comprising selecting the two or more images of the plurality of images based on orientation of the observed face.

12. The computer-implemented method of claim 11, wherein the selection of the two or more images of the plurality of images is further based on a facial feature of interest.

13. The computer-implemented method of claim 11, wherein the selection of the two or more images of the plurality of images is further based on facial orientation of the facial data of the database.

14. The computer-implemented method of claim 11, further comprising displaying on a graphical display data indicative of the one or more selected identities.

15. The computer-implemented method of claim 14, further comprising prompting a user to confirm the selection of the two or more selected identities;

and

updating the database in response to the confirmation of the user.

16. A facial recognition apparatus, comprising:

an image receiver module including instructions that configure a processor to receive a plurality of images of an observed face for identification;

a template module including instructions that configure the processor to generate a template of the observed face based at least on two or more selected images of the plurality of images;

a search module including instructions that configure the processor to search a database of identities and corresponding facial data of known persons, the searching being based at least on the template of the observed face and the facial data of the known persons; and

a recognition module including instructions that configure the processor to select one or more identities of the known persons based at least on the search.

17. The facial recognition apparatus of claim 16, further comprising a communication module including instructions that configure the processor to transfer data indicative of the selected one or more identities of the known persons, wherein each identity of the database is associated with data that is based on one reference image.

18. The facial recognition apparatus of claim 16, further comprising an image sensor, wherein the image receiver module further comprises instruction that configure the processor to capture multiple images of an observed face from the image sensor.

19. A non-transitory, computer-readable medium storing instructions that, when executed, cause a processor to perform a method comprising of:

receiving a plurality of images of an observed face for identification;

generating a template of the observed face based at least on two or more selected images of the plurality of images;

searching a database of identities and corresponding facial data of known persons, the searching being based at least on the template of the observed face and the facial data of the known persons; and

selecting one or more identities of the known persons based at least on the search.

20. The non-transitory, computer readable medium of claim 19, wherein each identity of the database is associated with data that is based on one reference image, and wherein the one or more selected identities are associated with feature vectors within a threshold distance of the template of the observed face.

21. The non-transitory, computer readable medium of claim 19, wherein the method further comprises displaying on a graphical display data indicative of the two or more selected identities.