Google: here is the algorithm that reveals the missing details from the photos pixelated

It made him famous Rick Deckard in Blade Runner, but now you can really make thanks to Google. With neural networks, you can transform virtually indistinguishable images into detailed and clearly legible. Here’s how you can.

The AI ​​once again helping the man. Yes, because thanks to some neural networks, Google has managed to emulate what the famous cop Blade Runner Rick Deckard, did during Ridley Scott. Essentially thanks to some voice commands was possible, in the film, analyze a photograph in minute detail by zooming to capacity without losing image quality. Well, all this can now also be carried out, in reality, thanks to the algorithm that precisely the Mountain View company has achieved.

Google Brain is the department for artificial intelligence research and it is from them that the idea of ​​creating a program that would allow to make visible a color stain in a real detailed picture has emerged. The work behind all this is clearly enormous but its operation seems to be, at least in words, almost natural. Yes, because actually what happens is that in some way is made by the human brain.

Technically, we are faced with two neural networks that work first individually and then simultaneously. For they are made to observe a multitude of images that portray human faces and subsequently a neural network starts to compare an image pixelated (8×8 pixels) and thus a very low resolution with all those it has stored previously.

At this point, it comes in the second neural network that increases the resolution of 32×32 pixels and above all fits missing details. All this allows you to translate virtually unreadable images mug shots where you can recognize the face of this subject.

In short, the improvement and image recognition is undoubtedly an area in which technology will become the main protagonist and where just the giants like Google want to put the hand. Just the Mountain View company does not seem to be the first to engage in this field since the Facebook ad with Lumos, the platform for the recognition of images and videos.

It clear that the work is still long and these algorithms cannot take the place of humans, but it is clear how the artificial intelligence developments will allow good things about security and who knows maybe the levels of Blade Runner.