Have you ever wondered what a painting might look like if it had been a photograph? We suppose unless it is based on an actual location, we will probably have to use our imagination, but that might no longer be the case in the future. This is thanks to researchers at UC Berkeley (via PetaPixel) who have come up with an “Unpaired Image-to-Image Translation” technique.
Basically this technique uses AI/an algorithm in which it take a painting and transform it into a photograph. This is almost like a reverse-Primsa where the app turns photos into paintings, or images that resemble paintings thanks to AI-applied filters. As you can see in the image above, the results are pretty spot on and scarily realistic.
In fact not only can this technique turn paintings into photos, but it can transform objects from one thing to another. For example it can turn a horse into a zebra and vice versa, or maybe apples into oranges, or transform photos of a winter landscape into a summer scene, and so on. The best part of this technique is that it doesn’t actually need to have an example of the image they want to turn it into, meaning that it isn’t just copying a look and applying it to the image.
We’re not sure what the end goal is here for this particular technique, or if it is just to demonstrate an algorithm that could be applied to other things in the future, but as it stands it is pretty impressive.
Filed in . Read more about AI (Artificial Intelligence).