Structure Reference takes the trial and error out of image generation.
(Image credit: Adobe)
Adobe Firefly continues to evolve with today's introduction of a new feature designed to improve consistency when using generative AI to create variations of a reference image. Structure Reference is the latest attempt to make AI image generation more predictable and less random.
Following the introduction of style reference tools (and character reference in Midjourney), Adobe Firefly Structure Reference allows users to ask the AI to apply the structure of an existing image to new generations.
Adobe showed us that you can use Firefly's Structure Reference to redesign a room (Image credit: Adobe)
With Structure Reference applied, Adobe Firefly will create a structural template from a reference image and generate image variations with the same underlying composition. Demonstrating the tool at the annual Adobe Summit, Adobe showed how the tool could be used to redesign by uploading a picture or sketch of the existing room and hitting ‘generate’.
Another example showed how a children’s drawing could be turned into a striking neon-hued image while respecting the form of the original (see above). Paintings can be made to look photorealistic or vice versa. The tool can also be used to colorize black and white images and line drawings, and create a new subject that has nothing to do with the reference image but keeps the same general compositional structure.
Structure Reference can also be used to change the subject while retaining the underlying composition of an image (Image credit: Adobe)
"We’ve eliminated the trial and error process of having to write the perfect prompt to get an output that matches your imagination," Adobe said in blog post on the tool. Structure Reference can also be combined with the existing Style Reference feature to maintain consistency in both structure and style at the same time for greater creative control.
Adobe suggests that Structure Reference will allow creators to easily modify work, while marketers will be able to generate on-brand images to use across a campaigns, and designers easily create an image layout, as well as the structure and pose of a subject, without needing to find the perfect text prompt. It's another example of how the leading models are working to add more control and make AI image generation more practical for creative workflows (you can always leave the new feature off if you preferred the fun of the random extra limbs and chaotic compositions).
The New Structure Reference capabilities are available today in the Firefly web application for both free and paid subscribers (see below for subscription prices). The launch comes just days after Adobe added its first AI tools in Adobe Substance 3D, its suite of 3D programs.
Get the Creative Bloq Newsletter
Daily design news, reviews, how-tos and more, as picked by the editors.
It also comes a day after the announcement of Canva's acquisition of Affinity, one of Adobe's main competitors. Affinity's tools lack generative AI functions for now, but Canva does have text-to-image features.
Thank you for reading 5 articles this month* Join now for unlimited access
Joe is a regular freelance journalist and editor at Creative Bloq. He writes news, features and buying guides and keeps track of the best equipment and software for creatives, from video editing programs to monitors and accessories. A veteran news writer and photographer, he now works as a project manager at the London and Buenos Aires-based design, production and branding agency Hermana Creatives. There he manages a team of designers, photographers and video editors who specialise in producing visual content and design assets for the hospitality sector. He also dances Argentine tango.
