DragGAN
A research technique for editing images by dragging points on them, from the Max Planck Institute.
Highlights
- Intuitive Drag-and-Drop Interface: Simplifies the process of image creation for users of all proficiency levels.
- GAN Integration: Leverages the power of advanced Generative Adversarial Networks for precise image synthesis.
- High-quality Output: Guarantees detailed and accurate image results.
- Customization: Offers adjustable parameters to ensure desired output.
External link — opens vcai.mpi-inf.mpg.de in a new tab. DragGAN is a third-party product; we are not affiliated with it.
About DragGAN
What it is
DragGAN is a method for editing generated images through direct manipulation: you place points on an image and drag them, and the underlying generative adversarial network deforms the image so the content follows — changing a pose, a facial expression, the orientation of an object — while keeping it coherent. It came out of the Max Planck Institute for Informatics and attracted enormous attention on publication.
Why it's different
The interaction was the contribution, and it is genuinely important: before this, editing a generated image meant re-prompting and hoping, with no way to say keep everything and move that. Point-dragging gave precise spatial control over generative output for the first time. Two things to be clear about. This is research rather than a product — there is no polished application, and using it means working with code. And it is built on GANs, while the field moved decisively to diffusion models, so the technique is now most visible as an influence on the editing features inside other tools rather than as something you would adopt directly.
How people use it
It is used by researchers and by developers building editing interfaces, as a reference implementation of a control scheme. For anyone wanting the capability rather than the method, the practical route is a current image tool that has absorbed the idea — most serious editors now offer some form of direct manipulation over generated content. It remains worth knowing as the origin of that interaction.
Written by the n3os team. We are not affiliated with DragGAN.
This listing was written from public information, without DragGAN’s involvement. If you own it and something here is wrong — or you would rather not be listed at all — email us and we will correct or remove it.
Get the ones worth knowing about
We write one of these for every tool worth the trouble. Get the new ones, plus what we have found genuinely useful lately.
Your address goes to Buttondown, who send the emails on our behalf. One click unsubscribes, and the list is never sold or shared.