There is now a new way to save distorted and poor quality photos. Google’s new artificial intelligence system has the ability to restore poor quality images through enhancing its resolution 16 times than the original. Given that, they are now able to refine the resolution of pixelated images and make more recognizable.
According to Google’s research work entitled, Pixel Recursive Super Resolution, Google’s neural network is technology’s approach wherein it is like hallucinating the missing data and pixels in order to sharpen the image. It has been reported that the technology undergoes a series of training so that it will be able to learn typical facial features as well as nuances. They fed the system with countless training images.
The second part of the process required the neural system to focus on comparing 8×8 pixel images with various 32×32 pixel images. The two neural networks will work together to redraw the best guesses of what the original facial features should look like. Google’s new AI was able to show significant improvements. It was able to identify and draw accordingly which facial feature it is. With that, Google’s Artificial Intelligence system is able to synthesize realistic details to the image by enhancing its resolution.
There are a lot of tech companies are putting great focus on images and that includes Facebook. Facebook’s blog post shared that its computer vision called Lumos is built for image and video understanding.
“Advances in deep learning have allowed us to make big improvements in image classification — questions like “What is in the image?” and “Where are the objects?” are being answered by systems more accurately than ever. We've advanced this research by designing techniques that detect and segment the objects in a given image,” as stated on the blog.
Meanwhile, in other new related Google, Jobs & Hire reported that Google continuously fights against fake news by providing high-quality standard articles.
Watch the video below: