Google Labs is testing a new image generator called Whisk. That can combine three images into one: one for the subject, one for the style, and one for the scene.
Google Labs is experimenting with a new image generator, Whisk, that combines three images into one image. To do so, Google is using the image-generation model Imagen 3, as it previously did with its video generator Veo and in Google Docs.
Creative in a new way
Users can upload one image for the subject, another for the scene and one for the style. So you can select a photo of yourself as the subject, a sunny landscape as the scene and a watercolor for the final style. Finally, add a text prompt to clarify details in the generated image.
The model automatically generates a caption for the photo that can be modified to further point out the desired outcome. Google writes in a blog that Whisk focuses only on some key features of the photo. Thus, the result may differ from what is expected. Thus, the subject may differ in height, hair or skin color. You can update the text prompt afterward, though.
Whisk is only available in the United States for now. When Europe can get started with it, the tech giant is not sharing for now.
read also