Google experiments with a new image generator that remixes three images into one creation

Google Labs is testing a new image generator called Whisk, which allows users to create remixed photos by prompting with images instead of text. Whisk utilizes Google’s Imagen 3 model to combine three images: one for the subject, another for the scene, and one for the style. Users can select different elements, such as a personal photo, a futuristic landscape, and an anime style, to generate a new image. The model creates a detailed caption to guide the image generation process, and users can also add text prompts for more specific outcomes. However, the results may vary in terms of subject characteristics like height or skin tone.

The tool is currently available only to U.S. users at labs.google/whisk, and Google allows users to view and edit the underlying prompts at any time. Despite its innovative approach, Whisk's focus on a few key characteristics from each image means that the generated results might not always align with user expectations.

Key takeaways

Google Labs is testing a new image generator called Whisk, which allows users to remix photos by altering the subject, scene, and style.
Whisk uses Google’s image-generation model, Imagen 3, to combine three images for creating a new image.
Users can input text prompts to further define the desired outcome, but results may vary in characteristics like height, weight, hairstyle, or skin tone.
The Whisk experiment is currently only available to users in the U.S. at labs.google/whisk.

Google experiments with a new image generator that remixes three images into one creation | TechCrunch

Key takeaways

Discussion (0)