AndroGen: Open-Source Synthetic Data Generation for Automated Sperm Analysis

This article presents Androgen, a software tool for generating synthetic microscopic images of male reproductive cell samples from different species. AndroGen addresses the lack of large, diverse and well-annotated datasets needed to train and evaluate machine learning models in automated sperm analysis. The tool enables the creation of highly customizable synthetic datasets without relying on real images or generative training models, making it especially useful in data-scarce research environments. The software is open source and has been released under the AGPL-3.0 license. It features a user-friendly GUI and includes several predefined configurations. The architecture of the system is presented in detail, and its capabilities are evaluated through three case studies using quantitatively and qualitatively metrics. Experimental results show that Androgen is a powerful tool that allows generating complete datasets of synthetic microscopic sperm samples, with a significant possibilities of customisation.