Adobe has begun the widespread rollout of a new generative artificial intelligence feature in its graphic editing software Photoshop. This feature involves creating images based on text descriptions using the Firefly model. It was announced back in April this year.
Previously, the Generative Fill feature in Photoshop only allowed for adding, extending, or removing specific parts of an image. Now, you can use this function to create an image from scratch by entering a description into Photoshop's updated generative AI tool, similar to large language models like Dall-E and Mid Journey. The resulting image can then be further refined.
“This really speeds up the creation time,” said Erin Boys from Adobe in April. “The idea of transferring something from your mind to canvas has never been so simple.”
This feature is based on the Firefly Image 3 model, which recently sparked negative reactions from artists towards Adobe. Creators were outraged by wording in Adobe's recent terms of service, interpreting it as allowing the company to freely use their work to train its generative AI models. However, Adobe later clarified that it is “committed to building creator-friendly AI,” meaning it will “never train models on client content.”
In addition to image creation, Adobe has introduced the Enhance Detail function within Generative Fill. For Illustrator, the Generative Shape Fill feature has been made available, allowing for the addition of detailed vectors in a designer's unique style, Enhanced Text to Pattern (which creates customizable vector patterns in the artist's style), and Style Reference. The company has also added a Mockup tool for producing “high-quality visual prototypes of art on items such as product packaging,” along with enhanced selection capabilities and more.
Source: Engadget
Comments (0)
There are no comments for now