Google Research is working on DreamBooth, a new tool artificial intelligence that can generate selfies starting from a really existing image. The functionality would therefore allow to create infinite selfies from just one real selfie.
It is a technology that could completely revolutionize the world of content creators and influencers, who will no longer have to take the time to take pictures in various places and with different outfits.
“When photography came along, many feared it would spell the end of art because it threatened to disrupt important areas of work, such as landscape painting and portraiture,” he explained James Manyika, senior vice president at Google Research. “However, in many ways the opposite has turned out to be true. Freed from the need to accurately reproduce reality, painters have gone to new places, giving birth to impressionism, modernism and much more”.
How DreamBooth works, Google’s AI to generate selfies
DreamBooth is the result of a collaboration between Google Research and Boston University. It is one text-image generation tool which allows people to upload photos and then request the generation of more similar images. Robert Wongvice president of the Google Creative Lab, said his team used the technology to generate an infinite amount of content starting from a few images of itself.
Unlike other AI selfie tools, which produce unrealistic images such as AI lensDreamBooth produces absolutely realistic images, which can safely be mistaken for real photos.
In addition to DreamBooth, Google is developing several other AI-based tools. There is for example StyleDrop, which allows you to generate logos and icons for corporate branding. Or again AI-Replywhich generates draft responses to YouTube comments.