Meta Showcases Its AI-Based Image Generation Tool
Meta has showcased an image generation tool called Make-A-Scene that requires users to provide a text-based prompt as well as an initial sketch of what they want to create.
"Prior state-of-the-art AI systems that generated awe-inspiring images primarily used a text description as input," Meta says(Opens in a new window). "But text prompts, like 'a painting of a zebra riding a bike,' generate images with compositions that can be difficult to predict."
The company says this can prevent someone from feeling "a strong sense of pride and ownership over the content" they've asked the AI to create. Make-A-Scene is supposed to address that problem by giving people more control over what kind of art will be generated.
Remember the meme(Opens in a new window) about how drawing an owl can be broken down into two steps: drawing some circles and then drawing the rest of the owl? Make-A-Scene effectively allows people to say "owl," draw a few circles, and then watch as the AI-based tool draws the rest of the owl.
Meta says this research is part of its "commitment to exploring ways in which AI can empower creativity – whether that’s bringing your 2D sketches to life, using natural language among other modalities to create 3D objects, building entire virtual spaces, or any other creative project."
The company connects these efforts to the metaverse, of course, but it's not hard to imagine social platforms like Facebook and Instagram benefiting from this research as well. People like to share their art; a tool like Make-A-Scene could give them more opportunities to do just that.
Meta says it plans to present on Make-A-Scene at the ECCV 2022(Opens in a new window) conference in Tel Aviv in October. More information about the research
Read more on pcmag.com
