BAGAN: Data Augmentation with Balancing GAN

Abstract

Image classification datasets are often imbalanced, characteristic that negatively affects the accuracy of deep-learning classifiers. In this work we propose balancing GAN (BAGAN) as an augmentation tool to restore balance in imbalanced datasets. This is challenging because the few minority-class images may not be enough to train a GAN. We overcome this issue by including during the adversarial training all available images of majority and minority classes. The generative model learns useful features from majority classes and uses these to generate images for minority classes. We apply class conditioning in the latent space to drive the generation process towards a target class. The generator in the GAN is initialized with the encoder module of an autoencoder that enables us to learn an accurate class-conditioning in the latent space. We compare the proposed methodology with state-of-the-art GANs and demonstrate that BAGAN generates images of superior quality when trained with an imbalanced dataset.

References in corpus (10)

• The Effectiveness of Data Augmentation in Image Classification using Deep Learning
• Conditional Image Synthesis With Auxiliary Classifier GANs
• Improved Techniques for Training GANs
• Adversarial Feature Learning
• Adversarially Learned Inference
• VEEGAN: Reducing Mode Collapse in GANs using Implicit Variational Learning
• Unrolled Generative Adversarial Networks
• Plug & Play Generative Networks: Conditional Iterative Generation of Images in Latent Space
• Are GANs Created Equal? A Large-Scale Study
• DeLiGAN : Generative Adversarial Networks for Diverse and Limited Data

Cited by in corpus (42)

• A Comprehensive Survey of Image Augmentation Techniques for Deep Learning
• Ten Years of Generative Adversarial Nets (GANs): A survey of the state-of-the-art
• Selective Synthetic Augmentation with HistoGAN for Improved Histopathology Image Classification
• Enhanced Balancing GAN: Minority-class Image Generation
• Data Augmentation in High Dimensional Low Sample Size Setting Using a Geometry-Based Variational Autoencoder
• Balanced Semi-Supervised Generative Adversarial Network for Damage Assessment from Low-Data Imbalanced-Class Regime
• Gradient based sample selection for online continual learning
• Free-form tumor synthesis in computed tomography images via richer generative adversarial network
• Model Patching: Closing the Subgroup Performance Gap with Data Augmentation
• Fully Automatic Electrocardiogram Classification System based on Generative Adversarial Network with Auxiliary Classifier
• Transient-optimised real-bogus classification with Bayesian Convolutional Neural Networks -- sifting the GOTO candidate stream
• Positive-Unlabeled Learning with Adversarial Data Augmentation for Knowledge Graph Completion
• Generative Image Translation for Data Augmentation of Bone Lesion Pathology
• Distilling and Transferring Knowledge via cGAN-generated Samples for Image Classification and Regression
• Weather GAN: Multi-Domain Weather Translation Using Generative Adversarial Networks
• Regularization Methods for Generative Adversarial Networks: An Overview of Recent Studies
• Data Augmentation and Classification of Sea-Land Clutter for Over-the-Horizon Radar Using AC-VAEGAN
• Evaluating Generic Auto-ML Tools for Computational Pathology
• Neural Network Normal Estimation and Bathymetry Reconstruction from Sidescan Sonar
• Learning Stage-wise GANs for Whistle Extraction in Time-Frequency Spectrograms
• Class-Weighted Classification: Trade-offs and Robust Approaches
• Expansion of Cyber Attack Data From Unbalanced Datasets Using Generative Techniques
• Adversarial Classifier for Imbalanced Problems
• MixerGAN: An MLP-Based Architecture for Unpaired Image-to-Image Translation
• Imbalanced Data Learning by Minority Class Augmentation using Capsule Adversarial Networks
• Speech Driven Talking Face Generation from a Single Image and an Emotion Condition
• CMTS: Conditional Multiple Trajectory Synthesizer for Generating Safety-critical Driving Scenarios
• PMC-GANs: Generating Multi-Scale High-Quality Pedestrian with Multimodal Cascaded GANs
• GAN-based Data Augmentation for Chest X-ray Classification
• FLANNEL: Focal Loss Based Neural Network Ensemble for COVID-19 Detection
• Selective Synthetic Augmentation with Quality Assurance
• Cavity Filling: Pseudo-Feature Generation for Multi-Class Imbalanced Data Problems in Deep Learning
• Removing Class Imbalance using Polarity-GAN: An Uncertainty Sampling Approach
• Learning More with Less: Conditional PGGAN-based Data Augmentation for Brain Metastases Detection Using Highly-Rough Annotation on MR Images
• Memory-guided Unsupervised Image-to-image Translation
• Learning End-to-End Action Interaction by Paired-Embedding Data Augmentation
• Teaching a GAN What Not to Learn
• Multiclass Classification via Class-Weighted Nearest Neighbors
• Mining Domain Knowledge: Improved Framework towards Automatically Standardizing Anatomical Structure Nomenclature in Radiotherapy
• Towards Fair Cross-Domain Adaptation via Generative Learning
• Toward Zero-Shot Unsupervised Image-to-Image Translation
• GAN based ball screw drive picture database enlargement for failure classification