AI for image generation has become a trend. This was facilitated by the appearance of DALL-E 2, which is available to a limited number of testers, and Craiyon – a new version of DALL-E Mini, open to everyone.
Meta is also not far behind – the company showed its tool for generating images based on artificial intelligence. The developers hope that it will help build exciting worlds in the metaverse and create quality digital art.
Images created by Make-a-Scene. Source: Meta
The special feature of the new tool is that it gives more control over the output image. All its predecessors are based on a text phrase, which is then transformed into a picture in various ways. This was not enough for Meta.
“To realize AI’s potential to push creative expression forward, people should be able to shape and control the content a system generates,” says Mark Zuckerberg.
To generate an image Make-A-Scene uses not only text descriptions, but also sketches, created by users. This allows people to influence the composition of the picture. As a result, it outputs an image with a size of 2048 x 2048 pixels.
“It demonstrates how people can use both text and simple drawings to convey their vision with greater specificity, using a variety of elements, forms, arrangements, depth, compositions, and structures,” says Zuckerberg.
Now Meta has shared the technology with famous AI artists who will test its capabilities. It is not yet known when it will be fully open.
Excited to announce Make-A-Scene, our latest research tool Mark Zuckerberg just shared. Make-A-Scene is an exploratory concept that gives creative control to anyone, artists & non-artists alike to use both text & sketches to guide AI image generation: https://t.co/p9HNFy3VeY pic.twitter.com/Ir5U4IvikV
— Meta AI (@MetaAI) July 14, 2022