November 14, 2025 Dataset Open Access

Training Data and Models for the paper: Data-efficient U-Net for Segmentation of Carbide Microstructures in SEM Images of Steel Alloys

Chekhonin, Paul; Korten, Till; Gerçek, Alinda Ezgi; Hassan, Maleeha; Steinbach, Peter

This dataset contains scanning electron microscopy (SEM) images of steel alloys, including paired secondary electron (SE2) and in-lens (InLens) channels, with corresponding binary segmentation labels. The data supports full reproduction of results presented in the referenced manuscript.

Dataset Description

• Content: 13 pairs of SEM images of two reactor pressure vessel (RPV) steels:
  • JFL: IAEA reference RPV base metal steel
  • ANP-10: Western type RPV steel
• Acquisition:
  • JFL: Zeiss NVision 40 microscope
  • ANP-10: Zeiss Ultra 55 microscope
  • Both SE and InLens detectors used simultaneously.
• Resolution: 2048 × 1404 pixels per image
  • 2048 px width corresponds to 14.3 µm (JFL) or 11.5 µm (ANP-10).

Using the dataset to reproduce the results of the manuscript

Download the zip file into the data/ subdirectory of the code repository and extract the archive:

cd data/
unzip data.zip

Dataset Structure

These directories contain the relevant data for the manuscript:

cloud/
├-─ preprocessed/
│   ├── hold-out/
│   ├── images/
│   └── labels/
├── processed_tiles/
│   ├── images/
│   └── labels/
├── tb_logs/
│   ├── unet_model/

Preprocessed

pre-processed whole images and corresponding labels

Processed Tiles

tiled images and labels

tb_logs

trained model weights