Materials Data Science and Informatics (IAS-9)

This research project is dedicated to the precise semantic segmentation of nanoparticles in high-resolution transmission electron microscopy (TEM) images. By employing state-of-the-art deep learning techniques tailored for semantic segmentation tasks, we aim to accurately delineate the boundaries and identify the specific classes of nanoparticles within complex TEM images. This endeavor involves the development of novel neural network architectures and training strategies optimized for handling the unique challenges posed by high-resolution TEM data, including the presence of noise, variability in particle morphology, the scale of nanoparticles, and the lack of annotated datasets. We use the publicly available HRTEM image dataset (Sytwu et al., 2022) in this project.

Related Publications:

• Bashir Kazimi, Stefan Sandfeld, "Enhancing Semantic Segmentation in High-Resolution TEM Images: A Comparative Study of Batch Normalization and Instance Normalization", Microscopy and Microanalysis, Volume 31, Issue 1, February 2025, ozae093, https://doi.org/10.1093/mam/ozae093