Benchmark Dataset for Locating Atoms in STEM images
收藏资源简介:
本数据集名为‘Benchmark Dataset for Locating Atoms in STEM images’,由威斯康星大学麦迪逊分校材料科学与工程系的研究团队创建。数据集包含300张模拟和实验的扫描透射电子显微镜(STEM)图像,所有图像均标注了原子列位置。该数据集旨在通过一系列图像质量指标,评估不同晶格中原子定位模型的性能。数据集内容涵盖多种晶格类型、缺陷和界面,适用于研究原子级分辨率图像中的缺陷、固态反应动态、相变等问题。数据集通过Foundry服务提供,支持快速、便捷的模型测试和数据再利用。
This benchmark dataset, named *Benchmark Dataset for Locating Atoms in STEM images*, was developed by a research team from the Department of Materials Science and Engineering, University of Wisconsin-Madison. The dataset consists of 300 simulated and experimental scanning transmission electron microscopy (STEM) images, all of which are annotated with atomic column positions. This dataset is designed to evaluate the performance of atomic localization models across diverse lattices using a series of image quality metrics. It covers a wide range of lattice types, defects and interfaces, making it applicable to research on defects in atomic-resolution images, dynamics of solid-state reactions, phase transitions and other related topics. The dataset is available via the Foundry service, enabling fast and convenient model testing and data reuse.




