Assisted photo-tagging with facial recognition models
View Patent ↗Embodiments of the invention perform assisted tagging of images, including tagging of people, locations, and activities depicted in those images. A batch of images is received comprising images of faces, including at least some faces that have not yet been tagged. A facial recognition algorithm is applied to the faces to determine matching data comprising possible tags for each untagged face. A logic engine applies logic rules to reduce the likelihood that certain matches are correct. The most likely match from among the possible matches is selected for suggestion to the user for verification. Once verified, the metadata of the image indicating the recognized people within the image is updated.
1. A computer-implemented method for recognizing and tagging faces, comprising:
receiving a batch of images comprising faces by one or more computer processors;
identifying an untagged face in an image in the batch by the one or more computer processors;
matching the untagged face to a plurality of facial recognition models associated with known persons by the one or more computer processors, to produce matching data comprising possible tags for the untagged face, wherein each of the possible tags is associated with a known person and a probability factor; and
automatically selecting a provisional tag for the untagged face from the possible tags based on the associated probability factors, comprising:
identifying, by the one or more computer processors, a highest probability factor among the probability factors that associate known persons with multiple untagged faces;
assigning a first untagged face to a first known person by the one or more computer processors based on the highest probability factor; and
eliminating at least one of the possible tags from a second untagged face by the one or more computer processors based on the highest probability factor.
2. The computer-implemented method of claim 1 , further comprising:
determining, by the one or more computer processors, a location at which the image was taken, wherein the untagged face in the image is prohibited to be tagged to a known person that is known to be at a different location when the image was taken.
3. The computer-implemented method of claim 1 , further comprising:
determining, by the one or more computer processors, an age of the untagged face in the image, wherein the untagged face is prohibited to be tagged to a known person that has a different age from the age determined for the untagged face.
4. The computer-implemented method of claim 1 , wherein the untagged face in the image is prohibited to be tagged to a known person that is already tagged to another face in the image.
5. The computer-implemented method of claim 1 , wherein the step of eliminating at least one of the possible tags from a second untagged face is in response to assignment of the first untagged face to the first known person.
6. The computer-implemented method of claim 1 , further comprising:
sending a facial recognition request comprising the image comprising the untagged face to a facial recognition sever; and
receiving the matching data from the facial recognition server.
7. The computer-implemented method of claim 1 , further comprising:
determining a gender of the untagged face in the image by the one or more computer processors, wherein the untagged face is to be tagged to a known person that has the same gender as the gender determined for the untagged face.
8. The computer-implemented method of claim 1 , wherein the matching data further comprises at least one of an estimated age for the face or a predicted gender for the face.
9. The computer-implemented method of claim 1 , wherein the matching data further comprises an unknown person tag indicating that no match was found.
10. The computer-implemented method of claim 9 , further comprising:
determining importance of people tagged as unknown based on a frequency that faces of the people appear in a user's images.
11. The computer-implemented method of claim 10 , further comprising:
presenting images of unknown people for identification by a user in order of importance.
12. The computer-implemented method of claim 1 , further comprising:
sending a plurality of provisional tags and the corresponding untagged faces to a client device for display at the client device;
receiving an indication that one or more of the provisional tags are incorrect from the client device; and
replacing the one or more of the provisional tags that are indicated to be incorrect with a provisional unknown person tag.
13. The computer-implemented method of claim 12 , further comprising:
grouping a subset of faces with an unknown person tag into a group based on a similarity between matching data corresponding to each of the subset; and
sending a request to the client device for a user to provide an identity for the group and indicate any faces in the group that do not correspond to the provided identity.
14. The computer-implemented method of claim 12 , further comprising:
receiving, by the one or more computer processors, an input from the user that indicates at least one of the untagged faces that has been incorrectly tagged.
15. The computer-implemented method of claim 12 , further comprising:
verifying tags not indicated to be incorrect; and
updating at least one of facial recognition models based on the tags that are not indicated to be incorrect.
16. The computer-implemented method of claim 12 , further comprising:
updating at least one of facial recognition models based on the one or more of the provisional tags indicated to be incorrect.