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

2026-08-09T06:19:29Z https://rodare.hzdr.de/oai2d

oai:rodare.hzdr.de:4124 2025-11-14T15:20:32Z openaire_data openaire_data user-rodare

true 3.1 HZDR.RODARE 10.14278/rodare.4124 Chekhonin, Paul 0009-0005-5029-3061 Helmholtz Zentrum Dresden Rossendorf (HZDR) Korten, Till 0000-0002-2315-9247 Helmholtz Zentrum Dresden Rossendorf (HZDR) Gerçek, Alinda Ezgi Helmholtz Zentrum Dresden Rossendorf (HZDR) Hassan, Maleeha 0009-0000-7917-7025 Helmholtz Zentrum Dresden Rossendorf (HZDR) Steinbach, Peter 0000-0002-4974-230X Helmholtz Zentrum Dresden Rossendorf (HZDR) Training Data and Models for the paper: Data-efficient U-Net for Segmentation of Carbide Microstructures in SEM Images of Steel Alloys Rodare 2025 machine learning SEM steel carbide segmentation image processing 2025-11-14 https://rodare.hzdr.de/record/4124 https://www.hzdr.de/publications/Publ-42225 10.14278/rodare.4123 https://rodare.hzdr.de/communities/rodare Creative Commons Attribution 4.0 International Open Access 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 <ul> <li>Content: 13 pairs of SEM images of two reactor pressure vessel (RPV) steels: <ul> <li>JFL: IAEA reference RPV base metal steel</li> <li>ANP-10: Western type RPV steel</li> </ul> </li> <li>Acquisition: <ul> <li>JFL: Zeiss NVision 40 microscope</li> <li>ANP-10: Zeiss Ultra 55 microscope</li> <li>Both SE and InLens detectors used simultaneously.</li> </ul> </li> <li>Resolution: 2048 × 1404 pixels per image <ul> <li>2048 px width corresponds to 14.3 µm (JFL) or 11.5 µm (ANP-10).</li> </ul> </li> </ul> Using the dataset to reproduce the results of the manuscript Download the zip file into the <code>data/</code> subdirectory of the code repository and extract the archive: <pre><code class="language-bash">cd data/ unzip data.zip</code></pre> Dataset Structure These directories contain the relevant data for the manuscript: <code>cloud/</code> <code>├-─ preprocessed/</code> <code>│   ├── hold-out/</code> <code>│   ├── images/</code> <code>│   └── labels/</code> <code>├── processed_tiles/</code> <code>│   ├── images/</code> <code>│   └── labels/</code> <code>├── tb_logs/</code> <code>│   ├── unet_model/</code> Preprocessed pre-processed whole images and corresponding labels Processed Tiles tiled images and labels tb_logs trained model weights