Google announced Whisk, an artificial intelligence-based tool that allows you to generate images using other images as a query instead of long text formulations.
Image source: blog.google
With Whisk, you can upload images using sample images as your main theme, scene, or style. If desired, you can supplement them with text prompts; and if the necessary pictures are not at hand, the system will offer its own – probably also generated by AI. Having received the result, you can add it to your favorites or download it; or improve it by adding or editing the text query. Whisk is intended for “quick visual sketching rather than meticulous, pixel-perfect editing”; it can “miss,” Google admits, so it allows you to edit the original suggestions.
The service is based on the latest version of the Imagen 3 image generator, which Google DeepMind announced along with the Veo 2 video generator – a powerful competitor to OpenAI Sora is currently available only to a limited number of users of the experimental VideoFX platform.