Conditional Image Synthesis With Auxiliary Classifier GANs

TLDR

Synthesizing high‑resolution photorealistic images remains a long‑standing challenge in machine learning. The paper introduces new methods to improve the training of generative adversarial networks for image synthesis. The authors develop a label‑conditioned GAN variant that generates 128×128 images with global coherence and propose two new analyses to assess discriminability and diversity of class‑conditional samples. The analyses show that 128×128 samples convey class information absent in low‑resolution images, are more than twice as discriminable than resized 32×32 samples across 1,000 ImageNet classes, and that 84.7 % of classes exhibit diversity comparable to real ImageNet data.

Abstract

Synthesizing high resolution photorealistic images has been a long-standing challenge in machine learning. In this paper we introduce new methods for the improved training of generative adversarial networks (GANs) for image synthesis. We construct a variant of GANs employing label conditioning that results in 128x128 resolution image samples exhibiting global coherence. We expand on previous work for image quality assessment to provide two new analyses for assessing the discriminability and diversity of samples from class-conditional image synthesis models. These analyses demonstrate that high resolution samples provide class information not present in low resolution samples. Across 1000 ImageNet classes, 128x128 samples are more than twice as discriminable as artificially resized 32x32 samples. In addition, 84.7% of the classes have samples exhibiting diversity comparable to real ImageNet data.