Google Labs, Google’s experimental arm, is Test a new image generator called Whisk. This tool allows people to invite images instead of the text, which allows them to remix a photo by modifying the subject, the scene and the style.
Whisk uses the Google, Imagen 3 images generation model, to combine three images: one for the subject, another for the scene and one for the style. For example, you can select a photo of yourself as a subject, a futuristic landscape as a scene and an anime style for the final look.
The model automatically generates a detailed legend of your images, which is then used to guide imagen 3 in the creation of a photo remix. You can also enter text prompts to further define the desired result, including detailed descriptions such as “the subject is to go flywheel”.
Since Whisk focuses only on some key characteristics of each image, the company explains that the results may not always meet your expectations. For example, the subject generated could differ in height, weight, hairstyle or complexion. Google says you can display and modify the underlying prompts at any time.
The experience is currently only available for users based in the United States lab.google/whisk.