GAN: Interactive Point-based Operations on Generated Image Manifolds
The DragGAN artificial intelligence image editing model utilizes the latest AI technology to precisely control pixels and edit images through simple dragging and selecting of points.
This model is an open-source project, with the source code set to be released in June, and provides natural stretching and manipulation of objects, facial expressions, and image layouts. It also includes a masking feature for precise modifications of specific areas of an image.
Official introduction and demonstration: https://vcai.mpi-inf.mpg.de/projects/DragGAN/
Github: https://github.com/XingangPan/DragGAN