OpenAI has added a new outpointing feature to its DALL-E artificial intelligence model, which allows the system to generate new visual effects that push the boundaries of any given image.
The title image shows DALL-E using human prompting to “imagine” that it is behind the scenes of Johannes Vermeer’s “Girl with a Pearl Earring” portrait. At the same time, even from the limited information provided by the portrait, the system “finishes” objects in accordance with the style of Vermeer, imitating the shadows and highlights of the original.
“Paramount Pictures Outpainting” by lu9_st #dalle #dalle2 #aiart #aiartwork pic.twitter.com/B3t0lBSCgz
— Best Dalle2 AI Art ? (@Dalle2AI) September 1, 2022
The outpointing feature can be used to extend original content. Many DALL-E users have used it to see beyond known images. However, like other AI systems that transform text into an image, the model requires humans to describe new visual effects. Therefore, in some cases, extended images are made in the same style as the original works. And in other cases, it turns out the post-apocalyptic Mona Lisa.
Well
UI/UX Design for the evening
Discover the profession of a UI / UX designer in 5 months, learn at a free hour
REGISTER!
“Outpainting: an apocalyptic Mona Lisa” by tonidl1989 #dalle #dalle2 #aiart #aiartwork pic.twitter.com/puYVxjyFMm
— Best Dalle2 AI Art ? (@Dalle2AI) September 1, 2022
The following video shows how artist August Kamp expands the image of the Girl with a Pearl Earring painting a little at a time, sometimes reworking the result generated by the DALL-E neural network.
DALL E: Introducing Outpainting
Extend creativity and tell a bigger story with DALL-E images of any size
blog: https://t.co/dBRzCXWDYF pic.twitter.com/9sOyDMPEb7
— AK (@_akhaliq) August 31, 2022
Note that recently the DALL-E image generator has become available to a wide range of users .
Source: The Verge
