Imagine taking a really good picture that you're really proud of and you want to post it on your socials like Instagram and Facebook. And as you zoom in to make sure the picture you shot is not that blurry or anything, uh-oh. Too bad, your picture is still blurry.
But thanks to a group of researchers who have figured out how to transform those blurry faces into mirror-quality faces, you can finally rest easy. It's not a new phone or nor is it a new camera, but using artificial intelligence.
The research team from Duke University located in Durham, North Carolina managed to create an algorithm that is capable of "imagining" realistic-looking faces just from the blurry and barely recognizable photos of people.
This AI is said to be eight times more effective than previous methods.
Cynthia Rudin, the computer scientist who led the research said: "Never have super-resolution images been created at this resolution before with this much detail," taken from The Independent UK.
Take note that the images that are generated by the AI is plausibly real. Meaning it cannot identify people anymore from poor quality cameras.
However, the researchers mentioned that the AI can make "any number of uncannily lifelike possibilities" from a single poor-quality snap, each looking subtly like a different person.
Read Also: How Artificial Intelligence Is Disrupting the Lending Industry in 2020
How AI fix blurry images
Also known as the Photo Upsampling via Latent Space Exploration system was developed by Dr Rudin and her team improving the images' resolution by up to 64 times.
This is made possible by reverse engineering the image from high resolution images that look similar to the low resolution image when down scaled.
Some image-enhacing tools would often take an image of low quality and predict what extra pixels were needed by trying to match pixels in high-res what the algorithm has seen before making the image fuzzy and indistinct.
In order to resolve that, Newsweek took note that PULSE scours a mass of AI-generated examples of high-resolution faces and searches for those that "look as much as possible like the input image when shrunk down to the same size."
Now, people can even see the most minute details of a person's face from his eyebrows to his dry skin and even a old lady's wrinkles.
"Instead of starting with the low resolution image and slowly adding detail, PULSE traverses the high resolution natural image manifold, searching for images that downscale to the original low resolution image," Rudin added from her research paper.
Theoretically, the AI can be used on low resolution images ranging from medicine and microscopy, to astronomy and satellite imagery that can remove the words "blurry photo" for good.
The research will be presented virtually at the 2020 Conference on Computer Vision and Pattern Recognition (CVPR) this week from June 14 to June 19 will also include Sachit Menon, Alex Damian, McCourt Hu, and Nikhil Ravi.
Read Also: Artificial Intelligence Could Make Us Lazy
