You will have seen photos on Facebook, Google Plus or other social networks, which testify to the location where they were taken. The location tracking function uses geotags. But the future is coming…
Today, MIT Tech Review reported details about a new project led by Tobias Weyand, an artificial intelligence expert at Google. The project breaks down photos down to the pixel level and then tries to match those pixels against a vast library of images to determine where they fit best.
It's the digital equivalent of a jigsaw puzzle where you have a piece and you try to find its exact position. Only in this case, there are hundreds of millions of alternatives.
The sophisticated AI, called PlaNet, uses over 90 million images with geographic coordinates from around the globe. In a test with 2.3 million images, PlaNet was able to determine the country of origin 28.4% of the time and continent of origin 48% of the time.
Although the results are not perfect, the initial numbers are significantly better than the results of tests conducted with humans.
