Google’s PlaNet can guess the location of your images
Google’s PlaNet can guess the location of your images

Technology major Google has taken another notable step forward towards creating machines that can perform tasks better and faster than humans. Last week, Google revealed PlaNet, a neural network developed to locate the exact location where a picture was captured. According to the MIT Technology Review, the difference between this latest technology and other image location software is that it grabs information, learns as it comes across more and more pictures.

PlaNet isn’t flawless. Project leader Tobias Weyand told MIT Technology Review that the network can just localize 3.6% of the pictures at street-level accuracy and 10.1% at city-level accuracy. But, amazingly, it has already defeated humans in trial runs.

With the help of a website known as GeoGuessr, PlaNet was tested opposing to 10 well-traveled humans and it won 28 of 50 rounds.

According to the publication, Weyand said, “We think PlaNet has an advantage over humans because it has seen many more places than any human can ever visit and has learned subtle cues of different scenes that are even hard for a well-traveled human to distinguish”.

However, probably the most astonishing fact about PlaNet is that one day you could carry the device in your pocket.

Weyand said that their model occupies 377 MB, and can easily fit in to the memory of a smartphone.

Last year, Google’s overhauled Photos app was introduced with some awesome tricks, including it can organize your pictures by location and face identification, venues, landmarks and animals.

Generally, if you wish to know where a picture has been clicked, you can see its metadata, which mostly includes GPS coordinates, and that is the way Google Photos function. The technology giant published an academic paper last week showing they can teach a deep-learning machine the skills of finding the location of a photo without using this metadata.

Check out More news from Telecom Sector :: Pharmaceutical Sector :: Auto Sector :: Infrastructure :: Real Estate