Séminaire LGF – Abdelouahid Bentamou – 16 mai 2024

Intitulé

Generating 3D Microstructure Images for O2 Fuel Cell Electrode using GANs Enhanced with Minkowski Functional

Résumé

This presentation presents a method using Generative Adversarial Networks (GANs) and Minkowski functionals to generate realistic 3D microstructure images of O2 fuel cell electrodes, crucial for materials optimization. Minkowski functionals, capturing porosity and specific surface area, enhance GAN training by guiding the generator to accurately replicate intricate microstructure details. Experimental results validate the approach’s ability to generate authentic 3D microstructure images, contributing to fuel cell technology advancements.