Here's what some of the famous paintings could have looked like had they been bigger.
Neural networks enthusiasts Denis Shiryaev and Monkeyinlaw have presented a collection of famous paintings expanded by DALL-E 2, OpenAI's recently-released neural network capable of creating images from text input, making realistic edits to existing images, removing unwanted parts from an image, etc. The authors utilized DALL-E 2 in an unorthodox way, turning famous paintings by Leonardo da Vinci, Vasily Vereshchagin, Johannes Vermeer, and many others into bigger versions of themselves by adding new details with the AI.
And here are the results:
The Mona Lisa by Leonardo da Vinci
Girl with a Pearl Earring by Johannes Vermeer
The Apotheosis of War by Vasily Vereshchagin
Barge Haulers on the Volga by Ilya Repin
Morning in a Pine Forest by Ivan Shishkin and Konstantin Savitsky
The Ninth Wave by Ivan Aivazovsky
Girl with Peaches by Valentin Serov
The Swan Princess by Mikhail Vrubel
Pink Mountains by Nicholas Roerich
The Christ Pantocrator of St. Catherine's Monastery at Sinai, one of the oldest religious icons
You can learn more about the project here. Also, don't forget to join our new Reddit page, our new Telegram channel, follow us on Instagram and Twitter, where we are sharing breakdowns, the latest news, awesome artworks, and more.