Drag Your GAN: Interactive Point-based Manipulation on the Generative Image Manifold
The authors propose a method that allows users to "drag" the content of any GAN-generated images by clicking a few handle points and target points on the image. This flexible point-based manipulation enables control of many spatial attributes like pose, shape, expression, and layout across diverse object categories. The approach consists of two main components: a feature-based motion supervision that drives the handle point to move towards the target position, and a new point tracking approach that leverages the discriminative generator features to keep localizing the position of the handle points. The authors demonstrate the advantage of DragGAN over prior approaches in the tasks of image manipulation and point tracking.