GeMix: Conditional GAN-Based Mixup for Improved Medical Image Augmentation

TL;DR

GeMix introduces a class-conditional GAN-based augmentation method that produces realistic, label-aware images, improving medical image classification performance over traditional mixup techniques.

Contribution

We develop GeMix, a novel two-stage framework that replaces heuristic pixel-wise mixup with learned, label-aware image synthesis using class-conditional GANs for medical imaging.

Findings

Improves macro-F1 scores over traditional mixup across multiple backbones.

Reduces false negative rate in COVID-19 detection.

Provides a drop-in replacement for pixel-space mixup with better semantic fidelity.

Abstract

Mixup has become a popular augmentation strategy for image classification, yet its naive pixel-wise interpolation often produces unrealistic images that can hinder learning, particularly in high-stakes medical applications. We propose GeMix, a two-stage framework that replaces heuristic blending with a learned, label-aware interpolation powered by class-conditional GANs. First, a StyleGAN2-ADA generator is trained on the target dataset. During augmentation, we sample two label vectors from Dirichlet priors biased toward different classes and blend them via a Beta-distributed coefficient. Then, we condition the generator on this soft label to synthesize visually coherent images that lie along a continuous class manifold. We benchmark GeMix on the large-scale COVIDx-CT-3 dataset using three backbones (ResNet-50, ResNet-101, EfficientNet-B0). When combined with real data, our method increases macro-F1 over traditional mixup for all backbones, reducing the false negative rate for COVID-19 detection. GeMix is thus a drop-in replacement for pixel-space mixup, delivering stronger regularization and greater semantic fidelity, without disrupting existing training pipelines. We publicly release our code at https://github.com/hugocarlesso/GeMix to foster reproducibility and further research.