Method and apparatus for mitigating face aging errors when performing facial recognition

Patent metadata

Field	Value
Publication number	US-9542390-B2
Application number	US-201414266044-A
Country	US
Kind code	B2
Filing date	Apr 30, 2014
Priority date	Apr 30, 2014
Publication date	Jan 10, 2017
Grant date	Jan 10, 2017

Abstract

A computer implemented method and apparatus for mitigating face aging errors when performing facial recognition. The method comprises receiving an indication of a face that needs to be searched in an image set, where each image in the image set comprises a timestamp that identifies a creation date of the image, the creation date being in a continuum of successive time intervals; and identifying the indicated face in images taken in each time interval of a plurality of successive time intervals for the indicated face, wherein each face found in images taken in a previous successive time interval is used as a reference set for identifying the face in images taken in a next successive time interval.

First claim

The invention claimed is:

1. A computer implemented method comprising:
receiving an indication of a face to be searched in an image set, where each image in the image set comprises a timestamp that identifies a creation date of the image, the creation date being in a continuum of successive time intervals;
performing facial recognition to identify the indicated face in one or more images in a first time interval from a plurality of successive time intervals in the continuum using a positive reference set comprising one or more faces from a previous time interval that match the indicated face and a negative reference set comprising one or more faces from the previous time interval that do not match the indicated face;
replacing the one or more faces from images from the previous time interval in the positive reference set with one or more faces from images from the first time interval that match the indicated face;
replacing the one or more faces from images from the previous time interval in the negative reference set with one or more faces from images from the first time interval that do not match the indicated face; and
performing facial recognition to identify the indicated face in one or more images in a second time interval from the plurality of successive time intervals using the positive reference set and the negative reference set, the second time interval being successive to the first time interval.

2. The method of claim 1 , further comprising:
removing images of the image set that do not comprise a face; and
ordering images of the image set according to the timestamp of each image.

3. The method of claim 2 , wherein the successive time intervals are identified as repeating time periods of one of months, weeks of years.

4. The method of claim 1 further comprising:
performing facial recognition on the one or more images in the previous time interval using the indicated face as a reference, where each image in the previous time interval comprises a timestamp that is within the previous time interval;
storing one or more images from the previous time interval that include the one or more faces that match the indicated face in a resultant image set;
generating the positive reference set for the previous time interval, the positive reference set comprising the one or more faces from images from the previous time interval that match the indicated face; and
generating the negative reference set for the previous time interval, the negative reference set comprising the one or more faces from images from the previous time interval that do not match the indicated face.

5. The method of claim 4 , further comprising: determining that the one or more images in the second time interval do not include the indicated face;
retaining the one or more faces from images from the first time interval that match the indicated face in the positive reference set; and
replacing the one or more faces from images from the first time interval in the negative reference set with one or more faces from images from the second time interval that do not match the indicated face.

6. The method of claim 5 , further comprising:
performing facial recognition to identify the indicated face in one or more images in a plurality of image groups corresponding to a plurality of successive time intervals;
determining that a number of image groups from the plurality of image groups that do not have the indicated face in any images meets a predefined threshold; and
requesting a selection of a face in an image comprising a timestamp in a next time interval, the next time interval successive to a last searched time interval.

7. The method of claim 1 , wherein performing facial recognition for the second time interval comprises using one or more images in the positive reference set as reference to identify the indicated face in the one or more images of the second time interval, and using the one or more images in the negative reference set as reference to identify faces to exclude from possible matches of the indicated face in the one or more images in the second time interval.

8. An apparatus for mitigating face aging errors when performing facial recognition comprising:
at least one processor; and
a non-transitory computer readable storage medium comprising instructions that, when executed by the at least one processor, cause the apparatus to:
receive an indication of a face to be searched in an image set, where each image in the image set comprises a timestamp that identifies a creation date of the image, the creation date being in a continuum of successive time intervals;
perform facial recognition to identify the indicated face in one or more images in a first time interval from a plurality of successive time intervals in the continuum using a positive reference set comprising one or more faces from a previous time interval that match the indicated face and a negative reference set comprising one or more faces from the previous time interval that do not match the indicated face;
replace the one or more faces from images from the previous time interval in the positive reference set with one or more faces from images from the first time interval that match the indicated face;
replace the one or more faces from images from the previous time interval in the negative reference set with one or more faces from images from the first time interval that do not match the indicated face; and
perform facial recognition to identify the indicated face in one or more images in a second time interval from the plurality of successive time intervals using the positive reference set and the negative reference set, the second time interval being successive to the first time interval.

9. The apparatus of claim 8 , further comprising instructions that, when executed by the at least one processor, cause the apparatus to:
remove images of the image set that do not comprise a face; and
order images of the image set according to the timestamp of each image.

10. The apparatus of claim 8 , further comprising instructions that, when executed by the at least one processor, cause the apparatus to:
perform facial recognition on the one or more images in the previous time interval using the indicated face as a reference, where each image in the previous time interval comprises a timestamp that is within the previous time interval;
store one or more images from the previous time interval that include the one or more faces that match the indicated face in a resultant image set;
generate the positive reference set for the previous time interval, the positive reference set comprising the one or more faces from images from the previous time interval that match the indicated face; and
generate the negative reference set for the previous time interval, the negative reference set comprising the one or more faces from images from the previous time interval that do not match the indicated face.

11. The apparatus of claim 10 , further comprising instructions that, when executed by the at least one processor, cause the apparatus to:
determine that the one or more images in the second time interval do not include the indicated face;
retain the one or more faces from images from the first time interval that match the indicated face in the positive reference set; and
replace the one or more faces from images from the first time interval in the negative reference set with one or more faces from images from the second time interval that do not match the indicated face.

12. The apparatus of claim 8 , further comprising instructions that, when executed by the at least one processor, cause the apparatus to:
perform facial re

Assignees

• Adobe Systems Inc

Inventors

• Shankaranarayana Sudhir Tubegere

Classifications

• G06V40/172 Primary

  Classification, e.g. identification · CPC title

• G06F16/00

  Information retrieval; Database structures therefor; File system structures therefor · CPC title

• G06F18/22

  Matching criteria, e.g. proximity measures · CPC title

• G06K9/6807

  Physics · mapped topic

• G06F17/30259

  Physics · mapped topic