Google Photos base of photos has grown to 13.7 petabytes

For the year, people made 24 billion selfies.

^{Composite images that correspond to optimal stimuli for a neuron-classifier of a cat and a person in Google’s neural network during the first experiment of 2012.}

Humanity continues to be heavily photographed, supplying Google with an abundant flow of information for training artificial intelligence systems. The neural network Google already knows how to identify a country from a photo , recognize objects , distribute photos into subject groups, automatically generate videos and collages, and much more.

Users themselves get free Google Photos hosting of unlimited size from Google, which is very convenient: you can free up space and never worry about losing photos.

Yesterday, Google Photos service summed up the first year of work . During this time, 13.7 petabytes of digital photos were removed from mobile devices and users' computers , the official blog says. Removing old photos from the phone, if there are copies in the cloud, was implemented in November 2015 in the version of the mobile application Google Photos 1.9.

A huge database of user photos is a great test for Google's neural network . Over the year, the program has put 2 trillion tags on the photos, including 24 billion self-portrait tags (selfies).

Google recognizes various objects in photographs, so that an archive search is possible by describing objects in the frame, and even using emoji: and others



Google Photos provides unlimited space for storing photos in "good quality", that is, subject to almost imperceptible image compression. The photo in the "original quality" is limited. Recoding is done automatically.

Another useful feature of Google Photos is collaborative albums that send notifications if a relative / friend adds a new photo to the album.


^{The list of hot keys is displayed by pressing Shift +?}

“Thank you for the wonderful first year - keep up the good work; all these selfie not appear by themselves! "- appealed to the users of Google Vice President, Google project manager Anil Sabharwal Photos (Anil Sabharwal).

In the near future, the neural network will become even smarter. For example, Anil Sabharwal promises that she will learn to accurately select the most beautiful pictures (hero shot). For example, in the album of wedding photos she will find the best shot of the newlyweds, where they look at each other.