Facebook has started rolling out an update to its photo tagging system that will now scan newly uploaded photos and alert all the users it recognizes in that image by default.
"These new features help you find photos that you’re not tagged in and help you detect when others might be attempting to use your image as their profile picture," said Joaquin Quiñonero Candela, Director, Applied Machine Learning at Facebook.
The Facebook photo tagging system works by scanning each image users upload on the site, looking for human faces. The system detects faces and computes a string of numbers (called template)
for each face.
It then compares this template to the face templates of other Facebook users and uses the results as friend suggestions for the photo tagger system. If a user wishes to tag a friend in the photo, that friend will receive a notification.
The new system that Facebook has started rolling out will send notifications to all users found in newly uploaded photos, not just those that have been actively tagged.
Facebook hopes this new tool will help curb abuse, bots, and account impersonators.
A variation of this facial recognition system will also be deployed for users with visual impairments as well. Facebook will use the same facial recognition scanner to detect people not tagged in an image and inform the user who's in the photo.
The system will roll out to all countries except Canada and the EU, where the company has not deployed its photo-tagging mechanism —a basic facial recognition algorithm— due to local user privacy laws.
The state of this new facial recognition system will be synced to the photo tag suggestions feature. Facebook says that if a user has photo tagging notifications disabled, he won't receive notifications when others upload photos of the user.
The user must visit the Facebook settings page and change the status of the new Face Recognition option to his liking.