Meta introduced Make-A-Scene, an AI research tool that uses sketches and text to create new artwork. Shared by Mark Zuckerberg, the tool is designed to give creative control to anyone, artists and non-artists alike.
Most AI systems rely solely on text prompts to generate images, but due to the limitations of text, it is often difficult to predict how the image will turn out.
Meanwhile, Make-A-Scene allows users to “communicate their vision with greater specificity using different elements, shapes, arrangements, depth, compositions and structures,” according to a post on Meta’s AI blog.
“This multi-modal generative AI method puts creative control in the hands of the people who use it, allowing them to describe and illustrate their vision through both textual descriptions and free-form sketches.”
The system works by learning important sketch components that are likely to be relevant to its creator, such as objects and animals. From there, it evaluates the quality of images created by different generative models. Make-A-Scene can still generate an image with only text prompts if the creator so chooses.
Meta noted that the AI was trained using millions of example images from public datasets. As such, “bias reflected in the training data affects the results of these models.”
“The AI industry is still at the beginning of understanding and addressing these challenges, and there is still much work to be done,” Meta said.
Excited to announce Make-A-Scene, our newest research tool that Mark Zuckerberg just shared. Make-A-Scene is an exploratory concept that gives creative control to anyone, artists and non-artists alike, to use both text and sketches to guide the generation of AI images: pic.twitter.com /Ir5U4IvikV
— Meta AI (@MetaAI) July 14, 2022
In other tech news, take a closer look at the Nothing Phone (1).
Add Comment