Image Augmentations for GAN Training

2020/06/04 by Zhengli Zhao, Zhao, Zhengli, Zizhao Zhang +7 · 10 citations
Computer Science · Engineering · Mathematics · #Advanced Neural Network Applications #Adversarial Robustness in Machine Learning #Computer Vision and Pattern Recognition (cs.CV) #FOS: Computer and information sciences #FOS: Electrical engineering #Generative Adversarial Networks and Image Synthesis #Image and Video Processing (eess.IV) #Machine Learning (cs.LG) #Machine Learning (stat.ML) #cs.CV #cs.LG #eess.IV #electronic engineering #information engineering #stat.ML

paper · pdf · doi:10.48550/arxiv.2006.02595

arxiv created 2020/06/04 · openalex publication_date 2020/06/04 · arxiv updated 2020/06/05 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28

Abstract

Data augmentations have been widely studied to improve the accuracy and robustness of classifiers. However, the potential of image augmentation in improving GAN models for image synthesis has not been thoroughly investigated in previous studies. In this work, we systematically study the effectiveness of various existing augmentation techniques for GAN training in a variety of settings. We provide insights and guidelines on how to augment images for both vanilla GANs and GANs with regularizations, improving the fidelity of the generated images substantially. Surprisingly, we find that vanilla GANs attain generation quality on par with recent state-of-the-art results if we use augmentations on both real and generated images. When this GAN training is combined with other augmentation-based regularization techniques, such as contrastive loss and consistency regularization, the augmentations further improve the quality of generated images. We provide new state-of-the-art results for conditional generation on CIFAR-10 with both consistency loss and contrastive loss as additional regularizations.

Citations

• ImageNet classification with deep convolutional neural networks
• Amortised MAP Inference for Image Super-resolution
• Faster R-CNN: Towards Real-Time Object Detection with Region Proposal\n Networks
• Hierarchical Implicit Models and Likelihood-Free Variational Inference
• Self-Attention Generative Adversarial Networks
• Large Scale GAN Training for High Fidelity Natural Image Synthesis
• High-Fidelity Image Generation With Fewer Labels
• Temporal Ensembling for Semi-Supervised Learning
• Unsupervised Data Augmentation for Consistency Training
• Unsupervised Representation Learning with Deep Convolutional Generative Adversarial Networks
• Self-training with Noisy Student improves ImageNet classification
• Momentum Contrast for Unsupervised Visual Representation Learning
• Adversarial AutoAugment
• FixMatch: Simplifying Semi-Supervised Learning with Consistency and Confidence
• A Simple Framework for Contrastive Learning of Visual Representations
• Learning Multiple Layers of Features from Tiny Images

Cited by

• A Turn Toward Better Alignment: Few-Shot Generative Adaptation with Equivariant Feature Rotation
• Deceive D: Adaptive Pseudo Augmentation for GAN Training with Limited Data
• Improved Consistency Regularization for GANs
• Cross-Modal Contrastive Learning for Text-to-Image Generation
• Differentiable Augmentation for Data-Efficient GAN Training
• Improving Augmentation and Evaluation Schemes for Semantic Image Synthesis
• Dual Contrastive Loss and Attention for GANs
• Functional Space Analysis of Local GAN Convergence
• PConv: simple yet effective convolutional layer for generative adversarial network
• One-Shot Generative Domain Adaptation
• Generative Adversarial Networks with Limited Data: A Survey and Benchmarking

Related

• On Data Augmentation for GAN Training
• Large Scale GAN Training for High Fidelity Natural Image Synthesis
• Differentiable Augmentation for Data-Efficient GAN Training
• Deceive D: Adaptive Pseudo Augmentation for GAN Training with Limited Data
• Instance-Conditioned GAN Data Augmentation for Representation Learning