This is pretty much the dream technology for cop shows

Feb 7, 2017 14:42 GMT  ·  By

Google Brain has now created a new software that will make the dreams of every CSI fan who’s annoyed the main technology used throughout the series doesn’t exist come true. Well, it didn't exist, as it does now because the new Google software can create detailed images from tiny, pixelated images.

Zoom and enhance has been the target of so many jokes over the years because everyone with a computer knows that it’s impossible to get a clear picture when zooming in on an average picture. You’d just get loads of larger pixels and not a clear image. Or better said, used to. This new pixel recursive super resolution technology has really cool-looking results, extracting loads of details from a few source pixels.

For instance, Google Brain presented some 8x8 pixel images which it then turned into some pretty clear photos where you can actually tell facial features apart.

What is this sorcery, you ask? Well, it’s Google combining two neural networks. The first one, the conditioning network, works to map the 8x8 pixel source image against other high-resolution images. Basically, it downsizes other high-res images to the same 8x8 size and tries to make a match on the features.

Then, the second network comes into play, called the prior network. This one uses an implementation of PixelCNN to add realistic, high-res details to that 8x8 source image. If the networks know that one particular pixel could be an eye, when you zoom in, you’ll see the shape of an eye there. Or an eyebrow, or a mouth, for instance.

The zooming progression
The zooming progression

Easily fooled

The technology was put to the test and it was quite successful against humans. Human observers were shown a high-resolution celebrity face vs. the upscaled image resulted from Google Brain. Ten percent of the time, they were fooled. When it comes to the bedroom images used by Google for the testing, 28 percent of humans were fooled by the computed image.

It should be kept in mind, however, that while the idea is nice, the resulting images are pretty much guess work. They’re not real because you really can’t just make an image appear out of nothing. You can, however, use an educated guess and get something that might be close to accurate. An image of a suspect in a break-in, for instance, would get more details, but you won’t be able to say “hey, I know that guy” with complete certainty.

Several examples of how the technology works
Several examples of how the technology works

Photo Gallery (3 Images)

Several examples of how the technology works
The zooming progressionSeveral examples of how the technology works
Open gallery