Several teams of academics and Google have presented, DragGAN, an AI that certainly represents the future of photo editing.
Researchers from the University of Pennsylvania, MIT, Saarbrücken Research Center for Visual, Max Planck Institute and Google published a paper to present DragGAN, an AI-based tool which, like Midjourney or Dall -E, uses generative adversarial networks to create or modify an image in a very realistic way. According to them, current approaches to manipulating and/or creating images “lack flexibility and precision.”
Their invention provides an interactive way to control GANs (for Generative Adversarial Networks) by dragging points to reach target positions. These manipulations make it possible to control the pose, form, expression and arrangement of the various objects, human or animal, with precision. Even better, DragGAN performs these operations in near real time, and even creates content that does not originally exist, to fill in the deformations generated by the distortion of the image.
DRAGGAN IS AN AI-POWERED EDITING SOFTWARE CREATED BY GOOGLE
Google, thanks to Artificial Intelligence, will allow everyone to retouch photos in a very instinctive way. An operation that could previously take several hours for experienced graphic designers is now just a click away. The demonstrations available on the scientists' website are impressive. We imagine that the leaders of Adobe follow this invention very closely, as it could harm the sales of their star editing software, Photoshop.
Graphic designers will have to adapt and integrate this new technology, if they do not want to be put out of work, too, by Artificial Intelligence. At the general public level, DragGAN has certainly piqued the curiosity of millions of Internet users. At the moment, the site where you can consult the paper of researchers is overwhelmed with visits, so much so that you sometimes have to wait for it to appear.
Comments
Post a Comment