Monday , March 12 2018

More of the new Photos app: Automatically recognises faces, cars, skylines and more


We took a look at Google’s new Photos app yesterday, and it has a tonne of new features. Another new addition is the intelligence behind the app, which it uses to recognise people, places and things.

Google’s Photos service already uses some smarts to create its auto-awesome stories and compilations, but that only goes so far. The new Photos app takes this a step further. Not only will it pick out and identify faces, but it will recognise objects too — cars, food, places, structures and more.

One can only imagine that the computing power required to make these identifications and matches implies that this is a cloud-powered feature, though it is plausible that it could be performed on device in future.

The new auto-detected things are shown in the search interface, accessible via a floating action button in the main screen. As Android Police have rightly picked up, this really isn’t the most intuitive place for it — it could be better done. Regardless, the feature works pretty well – it picks out faces even if they’re hiding more in the background of a photo.

“Things” isn’t quite so fine tuned, but it does seem to work rather well depending on the subjects of your photos. Some of the groups which have been identified thus far include cars, skylines, dogs, flowers, stadiums, weddings, ships, food and more.

The accuracy (or wit) with which Photos groups things will be interesting to explore; cats can be mixed up with dogs (depending on how the dog looks I guess!), ships can be interchangeable with spaceships, and so on. Whether Google gets it right or not, you are able to manually remove images from groupings if you like, though it’s unknown whether this will feed into Google’s server side algorithms to improve future matches.

The guys over at Android Police have clearly got some unbridled access to this app to have seen what it can do, and sadly we can only report their findings — no one else has this kind of access to the app thus far.

Source: Screenshots from Ryan @ Android Police.
Via: Android Police.

Chris Rowland   Editor and Publisher

Chris has been at the forefront of smartphone reporting in Australia since smartphones were a thing, and has used mobile phones since they came with giant lead-acid batteries that were "transportable" and were carried in a shoulder bag. He saw the transition from AMPS to GSM, loved the Motorola StarTac, and got into Palm technologies in a big way. The arrival some years later of the original iPhone, and then the early Androids, awoke a new interest in mobile technology, and Chris has been writing about it since.

Today, Chris publishes one of Australia's most popular technology websites, Ausdroid. His interests include mobile (of course), as well as connected technology and how it can make all our lives easier.

Join the Ausdroid Conversation

5 Comments on "More of the new Photos app: Automatically recognises faces, cars, skylines and more"

newest oldest most voted
Valued Guest

Hopefully they fix the app so that when you choose to look at images “On device” it will show nice, high quality images, not the poor pixelated images it currently shows.

For whatever reason, photos look horrible when viewing them on the Google Photos app rather than the default gallery app.

Ausdroid Reader

There is a bit of spookiness to this seemingly cute bit of tech.
The Google AI is learning to recognize and understand the context of more Things. It’s brain is getting smarter.

Valued Guest

Absolutely. It won’t be long until Skynet really does have us.

Ausdroid Reader

I love it when intake photos of serial numbers on the back of equipment, and it makes an Autoawesome animation of it. Or a dramatic presentation with music.

Ausdroid Reader

I’d have a very large dogs section in my Photos app. xD

Check Also

Exclusive: Optus trialing Unleashed plans, with unlimited mobile data included for $60 per month

Optus has been secretly trialing a number of unlimited data plans in Australia, in what …