Beyond Data Scarcity Optimizing R3GAN for Medical Image Generation from Small Datasets

TL;DR

This paper presents optimized training strategies for R3GAN to generate realistic medical images from small datasets, improving class balance and classification metrics in embryo imaging.

Contribution

It introduces specific training techniques for R3GAN tailored to small medical datasets, enhancing image realism and diagnostic utility.

Findings

Improved classification recall and F1-score for the three-cell class.

Effective data augmentation using R3GAN balances class distribution.

Enhanced model robustness in small-scale medical imaging tasks.

Abstract

Medical image datasets frequently exhibit significant class imbalance, a challenge that is further amplified by the inherently limited sample sizes that characterize clinical imaging data. Using human embryo time-lapse imaging (TLI) as a case study, this work investigates how generative adversarial networks (GANs) can be optimized for small datasets to generate realistic and diagnostically meaningful images. Based on systematic experiments with R3GAN, we established effective training strategies and designed an optimized configuration for 256x256-resolution datasets, featuring a full burn-in phase and a low, gradually increasing gamma range (5 to 40). The generated samples were used to balance an imbalanced embryo dataset, leading to substantial improvement in classification performance. The recall and F1-score of the three-cell (t3) class increased from 0.06 to 0.69 and from 0.11 to 0.60, respectively, without compromising the performance of other classes. These results demonstrate that tailored R3GAN training strategies can effectively alleviate data scarcity and improve model robustness in small-scale medical imaging tasks.