Dresden 2026 – scientific programme

Parts | Days | Selection | Search | Updates | Downloads | Help

AKPIK: Arbeitskreis Physik, moderne Informationstechnologie und Künstliche Intelligenz

AKPIK 1: Tutorial: Machine Learning Use Cases in Materials Science (joint session AKPIK/TUT)

AKPIK 1.3: Tutorial

Sunday, March 8, 2026, 16:50–17:30, HSZ/0003

Machine Learning-based Analysis of Electron Microscopy Images: Preprocessing and Synthetic Data Generation — •Amir Omidvarnia — Forschungszentrum Jülich, Jülich, Germany

In this tutorial session, participants will learn how to prepare electron microscopy (EM) images for machine learning (ML) analysis using a series of Jupyter Notebook demonstrations. The tutorial illustrates essential preprocessing steps such as denoising, normalization, and contrast enhancement using Python. The session then transitions to synthetic EM data generation, showing how classical augmentation and modern generative models can create controlled datasets that mimic real data. By observing these live examples, attendees will gain a conceptual understanding of how preprocessing pipelines and synthetic data strategies can be used for ML-based EM analysis.

Keywords: Image processing; Machine learning; Preprocessing; Sunthetic data generation; Electron microscopy