Robust Conditional Generative Adversarial Networks

TL;DR

RoCGAN introduces a robust conditional GAN model that uses an unsupervised pathway to improve reliability and performance in noisy environments, outperforming existing cGANs in image generation tasks.

Contribution

The paper proposes RoCGAN, a novel cGAN architecture that incorporates an unsupervised pathway to enhance robustness against noise, with theoretical and empirical validation.

Findings

RoCGAN outperforms state-of-the-art cGANs in natural scene and face image generation.

The model maintains stability and accuracy even with intense noise.

Theoretical analysis shows RoCGAN shares properties with traditional GANs.

Abstract

Conditional generative adversarial networks (cGAN) have led to large improvements in the task of conditional image generation, which lies at the heart of computer vision. The major focus so far has been on performance improvement, while there has been little effort in making cGAN more robust to noise. The regression (of the generator) might lead to arbitrarily large errors in the output, which makes cGAN unreliable for real-world applications. In this work, we introduce a novel conditional GAN model, called RoCGAN, which leverages structure in the target space of the model to address the issue. Our model augments the generator with an unsupervised pathway, which promotes the outputs of the generator to span the target manifold even in the presence of intense noise. We prove that RoCGAN share similar theoretical properties as GAN and experimentally verify that our model outperforms existing state-of-the-art cGAN architectures by a large margin in a variety of domains including images from natural scenes and faces.