Researchers from Google, MIT, and the University of Pennsylvania have developed DragGAN, a photo editing app that uses AI to manipulate images by dragging two main components. The AI tool regenerates areas of the image that did not exist before, allowing for precise control over the pose, shape, expression, and layout of diverse categories such as animals, cars, humans, and landscapes. The approach explores a new way of manipulating generative adversarial networks (GANs) and shows a lot of promise, with the ability to hallucinate occluded content and deform following the object's rigidity.
DragGAN is a new AI app that allows users to easily adjust photos and art by dragging across the image. The app uses GANs to manipulate images while maintaining their integrity, allowing users to deform images using interactive points and the generative AI assists with the changes for a smooth and accurate end result. The app currently works on 2D images but the team plans on releasing a version that works with 3D models.
Researchers have developed DragGAN, an AI-powered point-based image manipulation system that allows users to precisely control the pose, shape, expression, and layout of objects. DragGAN uses generative adversarial networks (GANs) to generate brand-new pixels in response to user input, enabling users to manipulate the pose, shape, expression, and layout of images with precise control over where pixels go. The user interface is straightforward, making it accessible to almost any user without understanding the underlying technology. The technology opens new directions for powerful image editing using generative priors.
