Google’s New AI System Unscrambles Pixelated Images
February 8th, 2017Via: Guardian:
Google’s neural networks have achieved the dream of CSI viewers everywhere: the company has revealed a new AI system capable of “enhancing” an eight-pixel square image, increasing the resolution 16-fold and effectively restoring lost data.
The neural network could be used to increase the resolution of blurred or pixelated faces, in a way previously thought impossible; a similar system was demonstrated for enhancing images of bedrooms, again creating a 32×32 pixel image from an 8×8 one.
I call bullshit on this. You see this stuff in Holllywood movies (“enhance, enhance”) and magically the details come out. I’ve worked with digital images for more than 20 years and can assure you that is impossible. Yes, you can increase the resolution of images, but that can’t “effectively restore lost data”. For example, in a photo with low enough resolution that you can make out the pixels, but little detail, you can increase resolution but the image just gets blurry. Unless the detail is there to begin with and you are looking at a low res preview, there is no chance to see what was never captured in the first place.