After launching its virtual assistant Meta AI for its social network, Meta introduced another AI-based tool called Imagine to generate images.
According to Android Authority, Imagine is a standalone service that users can use on their PC or smartphone to create unique images that perfectly reflect the user’s thoughts. It is identical to the Reimagine feature that Meta just deployed on Messenger and Instagram, but Imagine is an independent service, meaning it is “open” to all platforms.
Imagine is based on Emu technology, Meta’s image generation model, and is a tool based on AI image generation technology from text. It will be a competitor to imaging models such as the DALL-E or Midjourney. Its advantage is that it is easy to use and is freely available to everyone.
Imagina’s interface is quite simple. Users will find a text box where they can express their ideas and a “Generate” button so the AI can illustrate the text placed above. Meta will add watermarks to all images generated from Imagine to promote transparency and tracking of all images generated using AI models.
The watermark will be applied with the deep learning model and can be detected by the corresponding model. It can also recover common image manipulations such as cropping, changing colors, taking screenshots, etc. Meta plans to include this watermark in products that provide AI-generated images.
Like Meta AI, Imagine is currently only available to users in the US and it is unclear what Meta’s plans are for rolling this feature out to users globally.
