Did it ever occur to you while watching CSI or Hawaii Five-O how the cops zoom in for some non-existent details in photos? Seeing all those unnoticeable details in photos to pop out is quite amusing and one day, we could actually achieve something so impressive with artificial intelligence.
Researchers have been using machine-learning technologies and neural networks to fill in the missing details in photos and interestingly, there’s a website that’s already providing this magic to the people like you and me.
Let’s Enhance is an online experience that provides a simple drag-and-drop interface to enhance the quality of your small photos. The pictures uploaded on the site can be enlarged to 4 times and I think it’s pretty impressive.
It all starts with uploading a picture to their website and then clicking the “magic button.
The end results are pretty impressive.
Let’s Enhance restore the smoothness of the lines and even reproduces the reflections.
Founders Vladislav Pranskevičius and Alex Savsunenko told PetaPixel that Let’s Enhance uses machine-learning frameworks like EDSR and SRGAN.
SRGAN uses Generative Adversarial Network which is basically, pitting two neural networks against together repeatedly. Well, this is how the fine-tuning of an image is done by Let’s Enhance.
Note: All content is original & written by NewsCrux Team.