Adobe has added a new tool called “Generative Fill” that adds new AI-generated content to a selected fill area of an image based on a text description prompt. This tool is powered by Adobe Firefly and utilizes a technique “inpainting” that’s also used in Stable Diffusion.
Photoshop now let’s you summon new objects and augment creations layer by layer.
In Photoshop beta, it is now possible to type in a text prompt and Firefly will “fill in” the selected area of the image with a synthesized image. Firefly has been trained to associate imagery with text descriptions based on millions images from Adobe’s stock library.
The Generative Fill feature is available to Creative Cloud members with a Photoshop subscription in the Photoshop beta app. Adobe is planning to roll this feature out to all Photoshop users by the end of the year.