Google's Neural Network Guesses Image Locations With "Superhuman" Ease

[HardOCP News]

Screw using a phone book to track down Sarah Connor, the next Terminator will just scan a bunch of pictures using Google's new neural network. Based on your pupil dilation, skin temperature, and motor functions, I calculate an 83% probability that you will be located and killed because of your stupid geotagged Facebook photos.


The results make for interesting reading. To measure the accuracy of their machine, they fed it 2.3 million geotagged images from Flickr to see whether it could correctly determine their location. “PlaNet is able to localize 3.6 percent of the images at street-level accuracy and 10.1 percent at city-level accuracy,” say Weyand and co. What’s more, the machine determines the country of origin in a further 28.4 percent of the photos and the continent in 48.0 percent of them.

 

[Monkey God]

So if that is what is out tot he Public, then big brother has it and it's prob been used for a while now and much more advanced then googles....
LOL it's always 1984 in the good ole USA

More look home of big brother UK with 5 CCTV's on every corner.

 

[M76]

Guess I'm better than their machine, as I was able to pinpoint the continent, and the country of all 5 images geoguessr had fed me. Plus 4 out of 5 of them I've had it pinned within 1000 miles. I was only off once when I said Oahu, when it was in fact near San Diego. But that still counts as getting the continent and country right? If I take out the mass off water between the two I was actually very close :--D

 

[Inacurate]

I calculate an 83% probability that you will be located and killed because of your stupid geotagged Facebook photos.

I always said my lack of posting any pictures of myself online would be a good thing, never did I think I think it would save me from the terminators!

Huh, what is this trophie thing..

 

[viscountalpha]

Moral of the story, don't take dumb photos and put it on Flickr