Researchers from Max Planck Institute of Informatics, MIT and Google have recently developed DragGAN, a new AI app that enables one to easily adjust photos and art by dragging across the image.
This user-friendly tool stands on its own in comparison to other AI image tools like Dall-E and Midjourney. These image tools are capable of processing highly specific prompts but cannot output precise poses or layouts the way DragGAN is able to do.
The research team wrote: “Imagine being able to just “drag” any point of an image to exactly where you want it to be. That’s what we’re aiming to achieve with our new method, which we call DragGAN.”
According to the research team’s homepage, the abstract states that “Through DragGAN, anyone can deform an image with precise control over where pixels go, thus manipulating the pose, shape, expression, and layout of diverse categories such as animals, cars, humans, landscapes, etc. As these manipulations are performed on the learned generative image manifold of a GAN, they tend to produce realistic outputs even for challenging scenarios such as hallucinating occluded content and deforming shapes that consistently follow the object’s rigidity.”
Video shared by YouTube channel AI In A Minute:
Not only can you completely rotate a picture’s subject as if it were a 3D model, one can also manipulate the dimensions of a car, turn a smile into a frown, or change a face’s dimension, entirely. The research team’s homepage has been crashing due to the amount of traffic sent to the site from influencers on social media.
The GAN-based tool currently works on 2D images, but the team plans on releasing a version is better-equipped for 3D models.
Add comment