ParGAN: Learning Real Parametrizable Transformations
This addresses the need for more controllable image translation for researchers and practitioners, though it appears incremental as a generalization of cycle-consistent GANs.
The paper tackles the problem of controlling image-to-image transformations in GANs by proposing ParGAN, which learns transformations with intuitive parameter controls while preserving input content. The method achieves smooth interpolations and simultaneous learning of multiple transformations without requiring paired data.
Current methods for image-to-image translation produce compelling results, however, the applied transformation is difficult to control, since existing mechanisms are often limited and non-intuitive. We propose ParGAN, a generalization of the cycle-consistent GAN framework to learn image transformations with simple and intuitive controls. The proposed generator takes as input both an image and a parametrization of the transformation. We train this network to preserve the content of the input image while ensuring that the result is consistent with the given parametrization. Our approach does not require paired data and can learn transformations across several tasks and datasets. We show how, with disjoint image domains with no annotated parametrization, our framework can create smooth interpolations as well as learn multiple transformations simultaneously.