Google Neural Network Can Tell Location Of Any Photo

Posted: Mar 1 2016, 5:47am CST | by , in News | Latest Science News


Google Neural Network Can Tell Location of Any Photo
MIT Technology Review
  • Google’s Neural Network locates Regions where Photographs were Taken

Google’s neural network can locate the exact regions where photographs were taken. This ability is marvelous to say the least.

Usually if there is a landmark in the background of a pic, it can be easily identified by even the most thorough novice among us. However, were there no recognizable signs or monuments in the pic, then what would we do?

Worry not, since Google has applied its moonshot capabilities to come up with an answer to that question and it involves AI. An introduction to PlaNet would be helpful.

It is a neural network that is dependent upon image recognition technology to do the trick of identifying the locale on earth of various photos. 

Many clues such as building style, linguistic signs and the flora and fauna of the region play a role in this identification process. These little hints are matched with a collection of 126 million geotagged photographs.

These in turn are organized into 26000 grids. If you were to take a photo of a place in Brazil, the complex instrument could tell by the greenery and Portugese billboards that the locality is somewhere in South America.  

Even indoor photos could be identified according to the MIT Technology Review. Such extreme focal powers are the stuff of which legends are made of. It is all thanks to the miracles wrought by science. There is a small hitch though.

PlaNet isn’t a flawless work of perfection. It has its drawbacks. It is as good as the weakest link in its repertoire. Thus a photo that has very few signs to show the computer where to look for clues in the first place will not be identified except with some difficulty.

We must remember that the technology is still in its infancy. It will take some time for it to reach a certain level of impeccability.   

At the level of big city life, PlaNet could only pinpoint 10% of the images. It is better than human observation though. That is because it has greater identifying powers than many of the world’s biggest lovers of wanderlust.

Were this wonderful technology to be refined even further, such photos could be identified that are not recognizable even with a GPS turned on. Nevertheless, even at present, it is a vast improvement on our previous state of navigational know-how.

To create a gadget that can do things much faster and with greater efficiency than human beings is a delight and something to be lauded for its revolutionary implications. 

You May Like


The Author

<a href="/latest_stories/all/all/20" rel="author">Sumayah Aamir</a>
Sumayah Aamir (Google+) has deep experience in analyzing the latest trends.




Leave a Comment

Share this Story

Follow Us
Follow I4U News on Twitter
Follow I4U News on Facebook

Read the Latest from I4U News