Meta Unveils Make-A-Scene AI That Blends Text and Sketches Into Images
Meta introduces Make-A-Scene, an AI research tool that lets users generate digital art by combining text prompts with simple freeform sketches to control image layouts precisely.
Meta introduces a new artificial intelligence research concept called Make-A-Scene that allows users to generate digital images using both text prompts and simple freeform sketches. Unlike previous text-to-image AI systems that often produce unpredictable layouts, this new tool captures the scene layout so creators can dictate exactly where objects appear and how large they are.
The system gives users the flexibility to provide text-only prompts to generate an automatic layout, or to draw specific elements like animals or objects to guide the final artwork. Professional AI artists who tested the tool note that it allows them to quickly establish a base composition by hand, bridging the gap between traditional drawing and generative technology.
Beyond professional artists, Meta believes Make-A-Scene empowers anyone to express their creativity, as demonstrated by a program manager who uses the tool to turn his young children's playful sketches into vivid artwork. This research represents Meta's ongoing effort to build the next generation of creative AI tools that offer users more nuanced control over their digital creations.