Face recognition using face tracker classifier data
View Patent ↗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.
1. A method of in-camera face recognition training of a specific face within digital images acquired with a portable camera-enabled device, comprising:
using a lens, an image sensor and processor of a portable camera-enabled device to acquire digital images;
capturing or otherwise obtaining in the device multiple different images that include a face of a specific person:
identifying groups of pixels that correspond to the face within the multiple different images;
tracking the face within the multiple different images, wherein the tracking is performed in parallel with determining whether the identified face corresponds to the specific person;
selecting sets of classifiers as matching the faces identified in the multiple different images;
statistically analyzing the sets of classifiers to generate one or more reference classifier profiles of the face associated with the specific person, wherein the statistically analyzing comprises determining variance values for the sets of classifiers;
normalizing the reference classifier profiles to determine normalized face classifiers of an average face associated with the specific person;
generating a face recognition profile for the specific person based on the normalized face classifiers of the average face of the specific person; and
programming the processor such that when a probability that an identified face corresponds to the specific person associated with the face recognition profile exceeds a predetermined threshold, then the processor initiates a workflow, image processing or other pre-or post-image acquisition action on the portable camera-enabled device based on the recognition of the face as belonging to the specific person;
wherein the method is performed by one or more computing device.
2. The method of claim 1 , wherein the different images comprise two or more different poses of the face.
3. The method of claim 2 , wherein the different images comprise two or more different lighting conditions of the face.
4. The method of claim 1 , wherein the different images comprise two or more different lighting conditions of the face.
5. The method of claim 1 , wherein the one or more reference classifier profiles comprise multiple referenced classifier profiles, and the statistically analyzing comprises determining mean values for the sets of classifiers.
6. The method of claim 1 , further comprising tracking said face within a plurality of relatively low resolution images.
7. The method of claim 1 , wherein the selecting comprises aligning face region subwindows and determining probabilities for classifiers based upon which of the sets of classifiers is selected.
8. The method of claim 1 , further comprising determining probabilities for classifiers based upon the sets of classifiers that are selected.
9. A portable camera-enabled device capable of in-camera face recognition and face recognition training of a specific face within acquired digital images, comprising:
a lens and an image sensor to acquire digital images;
a processor programmed by processor-readable code embedded within one or more digital storage media, wherein the processor-readable code comprises:
a face detector component to program the processor to identify groups of pixels that correspond to a face within the digital images;
a face tracking component to program the processor to track the face within the multiple different images, wherein the tracking is performed in parallel with determining whether the identified face corresponds to a specific person;
a face recognition training component to program the processor to generate a face recognition profile for the specific person determined by programming the processor to perform the following:
selecting sets of classifiers as matching the faces identified in the multiple different images;
statistically analyzing the sets of classifiers to generate one or more reference classifier profiles of the face associated with the specific person, wherein the statistically analyzing comprises determining variance values for the sets of classifiers;
normalizing the reference classifier profiles to determine normalized face classifiers of an average face associated with the specific person;
generating the face recognition profile for the specific person based on the normalized face classifiers of the average face of the specific person; and
programming the processor such that when a probability that an identified face corresponds to the specific person associated with the face recognition profile exceeds a predetermined threshold, the processor initiates a workflow, image processing or other pre- or post-image acquisition action on the portable camera-enabled device based on the recognition of the face as belonging to the specific person.
10. The device of claim 9 , wherein the multiple different images comprise two or more different poses of the face.
11. The device of claim 10 , wherein the multiple different images comprise two or more different lighting conditions of the face.
12. The device of claim 9 , wherein the multiple different images comprise two or more different lighting conditions of the face.
13. The device of claim 9 , wherein the one or more reference classifier profiles comprise multiple reference classifier profiles, and the statistically analyzing comprises determining mean values for the sets of classifiers.
14. The device of claim 9 , wherein the selecting comprises aligning face region subwindows and determining probabilities for classifiers based upon which the sets of classifiers are selected.
15. The device of claim 9 , wherein the face tracking and recognition components are configured to operate simultaneously.
16. The device of claim 9 , wherein the processor-readable code is further configured for programming the processor to determine probabilities for classifiers based upon which the sets of classifiers are selected.
17. One or more non-transitory processor-readable storage media having code embedded therein for programming a processor to perform a method of in-camera face recognition training of a specific face within digital images acquired with a portable camera-enabled device, wherein the method comprises:
using a processor of a portable camera-enabled device to acquire digital images;
capturing or otherwise obtaining in the device multiple different images that include a face of a specific person;
identifying groups of pixels that correspond to the face within the multiple different images;
tracking the face within the multiple different images, wherein the tracking is performed in parallel with determining whether the identified face corresponds to the specific person;
selecting sets of classifiers as matching the faces identified in the multiple different images;
statistically analyzing the sets of classifiers to generate one or more reference classifier profiles of the face associated with the specific person, wherein the statistically analyzing comprises determining variance values for the sets of classifiers;
normalizing the reference classifier profiles to determine normalized face classifiers of an average face associated with the specific person;
generating a face recognition profile for the specific person based on the normalized face classifiers of the average face of the specific person; and
programming the processor such that when a probability that an identified face corresponds to the specific person associated with the face recognition profile exceeds a predetermined threshold, then the processor initiates a workflow, image processing or other pre- or post-image acquisition action on the portable camera-enabled device based on the recognition of the face as belonging to the specific person.
18. The one or more non-transitory processor-readable storage media of claim 17 , wherein the different images comprise two or more different poses of the face.
19. The one or more non-transitory processor-readable storage media of claim 18 , wherein the different images comprise two or more different lighting conditions of the face.
20. The one or more non-transitory processor-readable storage media of claim 17 , wherein the different images comprise two or more different lighting conditions of the face.
21. The one or more non-transitory processor-readable storage media of claim 17 , wherein the one or more reference classifier profiles comprise multiple reference classifier profiles, and the statistically analyzing comprises determining mean values for the sets of classifiers.
22. The one or more non-transitory processor-readable storage media of claim 17 , wherein the method further comprises tracking said face within a plurality of relatively low resolution images.
23. The one or more non-transitory processor-readable storage media of claim 17 , wherein the selecting comprises aligning face region subwindows and determining probabilities for classifiers based upon which the sets of classifiers are selected.
24. The one or more non-transitory processor-readable storage media of claim 23 , wherein the tracking is performed in parallel with determining whether the identified face corresponds to the specific person.
25. The one or more non-transitory processor-readable storage media of claim 17 , wherein the method further comprises determining probabilities for classifiers based upon which the sets of classifiers are selected.