Google Brain Makes It Possible to Zoom and Enhance Photos
If you’ve watched any detective show over the last decade or so, you would have at some point or another, seen a scene where the protagonists are faced with a grainy picture of a person’s face or the number plate on a car. The resident whizkid, and there’s always one, will zoom into the pic, do some keyboard magic and all of a sudden the grainy pixels magically turn into a legible string of numbers or a face that matches the mugshot the investigators have on file perfectly.
All this looks great, but it is pure fiction as anyone that has actually ever used Photoshop or GIMP will tell you: if an image doesn’t have enough pixels, no force on Earth will be able to make more of them. That is, until now: the whizkids at Google Brain have figured out a way to do it. Through the creative use of neural networks, the project is now able to take a grainy photo and add details to it. Though it won’t create a razor-sharp image like you see in TV shows, it’s still fairly impressive to see it in action.
Google Brain doesn’t of course conjure the pixels out of thin air, that law remains unbroken. Rather, it scans the image, then searches for ones like it on its massive database before filling in the details it thinks are missing. These “hallucinations” do the trick just fine, and who are we to argue with such impressive results?
Though the technology is still in its infant stage — and few police departments are going to be able to afford the supercomputer it will take to process these actions — it seems we are yet one step closer to making TV-show gimmicks a reality; hopefully Google’s next project will be the TARDIS.
Fergus has been tinkering with computers since he was a kid and likes to put a stop to parties by listing the specs of all the digital devices in the room. It's best not to let him near your computer since he'll take it apart and may not put it back together again before he leaves.