With the launch of its new object-aware image editing engineAdobe is ready to surprise photography enthusiasts and photo editing professionals. Project Stardust is based on generative artificial intelligence of Adobe FireFly Model 2 and promises to make editing people and objects in photos more intuitive, accessible and efficient than ever.
The news related to Project Stardust were previewed during the Adobe Max World Creativity Conference, which took place from 11 to 12 October. The annual event brings together Adobe engineers and researchers who offer a first look at ideas, new technologies and prototypes, which could become an integral part of the company’s products in the future.
Project Stardust removes objects from photos with content-aware fill
This new editing engine stands out for its ability to automatically locate complex elements in images, allowing users to manipulate them and therefore, move, edit or delete them effortlessly. The result is a surprisingly realistic modification, capable of blending the removed objects with the surrounding context.
Many will see similarities between Project Stardust and the feature Google Photos Magic Editor. Both try to simplify complex photo editsallowing users without professional-level skills to eliminate unwanted people and objects and create new content in the empty spaces resulting in realistic images that are completely consistent with the context.
The strength of Project Stardust lies in its simplicity of use. Several demonstrations show off how you can make changes to clothing or other items simply by writing what you want. This evokes the convenient contextual activity bar that we find in latest Photoshop updatesand which offers the most relevant options for your workflow.
Firefly AI powers Project Stardust
Stardust makes most of its own AI tools based on Adobe’s Firefly but giving them a new look. Part of the process includes making the Generative Fill, which involves its new Adobe FireFly Model 2: creates higher quality images with more precise human figures that don’t have the flaws of previous iterations. A new text-to-image model will supposedly create more details like pores or foliage, making them look more realistic.
Another feature is the AI’s ability to recognize key cultural landmarks and symbols. Besides that you can adjust various photo settings during creation, such as motion blur, field of view and other parameters. Another feature is Generative Matchwhich can apply the “style” of a separate image to a new photo created by AI “at scale.”
However, it is essential to clarify that, for the moment, this is what Adobe consider uno “sneak peek”: A public preview of some of the technologies they are developing, which may or may not be incorporated into a final product. Aya Philémon, project manager at Adobe, underlined that the innovations related to Project Stardust represent just a small portion of what’s coming. This suggests that Adobe is preparing many more surprises for users, which could revolutionize the way we interact with its products.
to know more: Artificial Intelligence: what it is and what it can do for us