Omni-GAN: On the Secrets of cGANs and Beyond
This work provides significant performance improvements in high-quality image generation and restoration for researchers and practitioners working with generative models, particularly in computer vision.
This paper introduces Omni-GAN, a variant of cGAN that addresses issues of unsatisfying performance and mode collapse in image generation. Omni-INR-GAN, a specific implementation, achieved new state-of-the-art Inception scores of 262.85 and 343.22 on ImageNet for 128x128 and 256x256 image sizes, respectively, surpassing previous records by over 100 points. It can also extrapolate low-resolution images to arbitrary resolutions, including over 60x higher resolution.
The conditional generative adversarial network (cGAN) is a powerful tool of generating high-quality images, but existing approaches mostly suffer unsatisfying performance or the risk of mode collapse. This paper presents Omni-GAN, a variant of cGAN that reveals the devil in designing a proper discriminator for training the model. The key is to ensure that the discriminator receives strong supervision to perceive the concepts and moderate regularization to avoid collapse. Omni-GAN is easily implemented and freely integrated with off-the-shelf encoding methods (e.g., implicit neural representation, INR). Experiments validate the superior performance of Omni-GAN and Omni-INR-GAN in a wide range of image generation and restoration tasks. In particular, Omni-INR-GAN sets new records on the ImageNet dataset with impressive Inception scores of 262.85 and 343.22 for the image sizes of 128 and 256, respectively, surpassing the previous records by 100+ points. Moreover, leveraging the generator prior, Omni-INR-GAN can extrapolate low-resolution images to arbitrary resolution, even up to x60+ higher resolution. Code is available.