AI-Generated Dataset of Sub-Visible Particles for Pharmaceutical Quality Control

This dataset contains high-fidelity synthetic images of sub-visible particles (SvPs) for silicone oil and air bubbles generated using a diffusion-based generative AI model, as described in Improved Sub-Visible Particle Classification in Flow Imaging Microscopy via Generative AI-Based Image Synthesis (https://arxiv.org/abs/2508.06021). The dataset addresses the severe class imbalance in flow imaging microscopy (FIM) data, where minority particle types such as silicone oil droplets and air bubbles are underrepresented compared to protein particles. Interested parties can generate an unlimited number of images using the trained diffusion models. For more information, see: https://github.com/utkuozbulak/svp-generative-ai