Face recognition using face tracker classifier data

• US 8,977,011 B2
• Filed: 10/23/2012
• Issued: 03/10/2015
• Est. Priority Date: 12/05/2008
• Status: Active Grant

• Current Assignee: FotoNation Limited (Adeia Inc.)
• Original Assignee: FotoNation Limited (Adeia Inc.)
• Inventors: Bigioi, Petronel, Corcoran, Peter
• Primary Examiner(s): Carter, Aaron W
• Application Number:
  US13/658,022
• Publication Number:
  US 20130044923A1
• Time in Patent Office: 868 Days
• Field of Search: 382115-127, 340/5.1, 340/5.2, 340/5.52, 340/5.53, 340/5.83, 356/71, 713/186, 902/3
• US Class Current: 382/118
• CPC Class Codes: G06V 40/161 Detection; Localisation; Normalisation
• CPC Class Codes: G06V 40/172 Classification, e.g. identification

Abstract

A method of determining face recognition profiles for a group persons includes determining with a multi-classifier face detector that a face region within a digital image has above a threshold probability of corresponding to a first person of the group, and recording probability scores which are analyzed for each classifier, including determining a mean and variance for each classifier for the first person. The process is repeated for one or more other persons of the group. A sub-set of classifiers is determined which best differentiates between the first person and the one or more other persons. The sub-set of classifiers is stored in association with the first person as a recognition profile.

Claims

1. A method of determining face recognition profiles for each of a group of at least two persons comprising analyzing a set of images where more than one face image of each person is available, comprising:

a) determining with a multi-classifier face detector that a face region within a digital image has above a threshold probability of corresponding to a first person of a group of persons;

b) recording a probability score for each classifier stage of the multi-classifier face detector, along with an overall probability score, including associating the probability scores with the first person;

c) repeating (a) and (b) for one or more further images each containing a face region that has above the threshold probability of corresponding to the first person;

d) analyzing the probability scores for each classifier and the overall probability score, including determining a mean and variance for each classifier for the first person;

e) repeating steps (a) through (d) for one or more further persons of the group of persons;

f) determining a sub-set of classifiers which best differentiates between the first person and the one or more further persons of the group; and

g) storing the sub-set of classifiers including associating said sub-set of classifiers with the first person as a recognition profile.

2. The method of claim 1, further comprising deriving the one or more further images each containing a face region of the first person from a set of preview stream images acquired within a same digital imaging device as the digital image.

3. The method of claim 1, further comprising storing the mean and variance for each classifier and the overall probability score, including associating these with the first person as a base recognition profile.

4. The method of claim 3, further comprising:

determining and storing base recognition profile data for the one or more further persons of the group of persons;

determining a new sub-set of classifiers which best differentiates the first person from the one or more further persons, and storing the new sub-set of classifiers as a new recognition profile for the first person, including associating the new sub-set of classifiers with the first person.

5. The method of claim 4, further comprising deriving multiple face images of the first person from a set of preview stream images, and determining the base recognition profile on these multiple face images from the set of preview stream images.

6. The method of claim 5, further comprising storing the base recognition profile of the first person and the base recognition profile data of the one or more further persons of the group of persons within a same digital imaging device used to capture the digital image and set of preview stream images.

7. A portable camera-enabled device capable of in-camera determination of face recognition profiles for each of a group of at least two persons when more than one face image of each person is available, comprising:

a lens;

an image sensor;

a processor; and

a processor readable medium having code embedded therein for programming the processor to perform a face recognition method that comprises;

(a) determining with a multi-classifier face detector that a face region within a digital image has above a threshold probability of corresponding to a first person of a group of persons;

(b) recording a probability score for each classifier stage of the multi-classifier face detector, along with an overall probability score, including associating the probability scores with the first person;

(c) repeating (a) and (b) for one or more further images each containing a face region that has above the threshold probability of corresponding to the first person;

(d) analyzing the probability scores for each classifier and the overall probability score, including determining a mean and variance for each classifier for the first person;

(e) repeating steps (a) through (d) for one or more further persons of the group of persons;

(f) determining a sub-set of classifiers which best differentiates between the first person and the one or more further persons of the group; and

(g) storing the sub-set of classifiers including associating said sub-set of classifiers with the first person as a recognition profile.

8. The device of claim 7, wherein the method further comprises deriving the one or more further images each containing a face region of the first person from a set of preview stream images acquired within a same digital imaging device as the digital image.

9. The device of claim 7, wherein the method further comprises storing the mean and variance for each classifier and the overall probability score, including associating these with the first person as a base recognition profile.

10. The device of claim 9, wherein the method further comprises:

determining and storing base recognition profile data for the one or more further persons of the group of persons;

determining a new sub-set of classifiers which best differentiates the first person from the one or more further persons, and storing the new sub-set of classifiers as a new recognition profile for the first person, including associating the new sub-set of classifiers with the first person.

11. The device of claim 10, wherein the method further comprises deriving multiple face images of the first person from a set of preview stream images, and determining the base recognition profile on these multiple face images from the set of preview stream images.

12. The device of claim 11, wherein the method further comprises storing the base recognition profile of the first person and the base recognition profile data of the one or more further persons of the group of persons within a same digital imaging device used to capture the digital image and set of preview stream images.

13. One or more non-transitory processor-readable storage media having code embedded therein for programming a processor to perform a method of in-camera determination of face recognition profiles for each of a group of at least two persons when more than one face image of each person is available, wherein the method comprises:

(a) determining with a multi-classifier face detector that a face region within a digital image has above a threshold probability of corresponding to a first person of a group of persons;

(b) recording a probability score for each classifier stage of the multi-classifier face detector, along with an overall probability score, including associating the probability scores with the first person;

(c) repeating (a) and (b) for one or more further images each containing a face region that has above the threshold probability of corresponding to the first person;

(d) analyzing the probability scores for each classifier and the overall probability score, including determining a mean and variance for each classifier for the first person;

(e) repeating steps (a) through (d) for one or more further persons of the group of persons;

(f) determining a sub-set of classifiers which best differentiates between the first person and the one or more further persons of the group; and

(g) storing the sub-set of classifiers including associating said sub-set of classifiers with the first person as a recognition profile.

14. The one or more non-transitory processor-readable storage media of claim 13, wherein the method further comprises deriving the one or more further images each containing a face region of the first person from a set of preview stream images acquired within a same digital imaging device as the digital image.

15. The one or more non-transitory processor-readable storage media of claim 13, wherein the method further comprises storing the mean and variance for each classifier and the overall probability score, including associating these with the first person as a base recognition profile.

16. The one or more non-transitory processor-readable storage media of claim 15, wherein the method further comprises:

determining and storing base recognition profile data for the one or more further persons of the group of persons;

determining a new sub-set of classifiers which best differentiates the first person from the one or more further persons, and storing the new sub-set of classifiers as a new recognition profile for the first person, including associating the new sub-set of classifiers with the first person.

17. The one or more non-transitory processor-readable storage media of claim 16, wherein the method further comprises deriving multiple face images of the first person from a set of preview stream images, and determining the base recognition profile on these multiple face images from the set of preview stream images.

18. The one or more non-transitory processor-readable storage media of claim 17, wherein the method further comprises storing the base recognition profile of the first person and the base recognition profile data of the one or more further persons of the group of persons within a same digital imaging device used to capture the digital image and set of preview stream images.