Informative sample generation using class aware generative adversarial networks for classification of chest Xrays
This addresses the challenge of class imbalance in medical imaging for disease detection, offering a method to reduce data requirements, though it appears incremental as it builds on existing active learning and GAN techniques.
The paper tackled the problem of training robust deep learning systems for disease detection from chest X-rays, which is hindered by limited images and severe class imbalance, by proposing an active learning framework that selects informative samples and uses a class-aware generative adversarial network for data augmentation, achieving state-of-the-art performance with about 35% of the full dataset.
Training robust deep learning (DL) systems for disease detection from medical images is challenging due to limited images covering different disease types and severity. The problem is especially acute, where there is a severe class imbalance. We propose an active learning (AL) framework to select most informative samples for training our model using a Bayesian neural network. Informative samples are then used within a novel class aware generative adversarial network (CAGAN) to generate realistic chest xray images for data augmentation by transferring characteristics from one class label to another. Experiments show our proposed AL framework is able to achieve state-of-the-art performance by using about $35\%$ of the full dataset, thus saving significant time and effort over conventional methods.