Enclosed we report all images and atomic positions used in the article: "Automated Image Analysis for Single-Atom Detection in Catalytic Materials by Transmission Electron Microscopy" The corresponding code can be found under: https://github.com/HPAI-BSC/AtomDetection_ACSTEM The folders are organized as follows: -ground truth data set (manually labeled images): - groundtruth_labeled_images_manual: images with manual labels in two formats (see .tif for pixel size) - groundtruth_atomic_positions_manual: tabulated atomic positions from manual labels - groundtruth_labeled_images_SAC-CNN: manual and SAC-CNN labeled images of ground truth set - groundtruth_atomic_positions_SAC-CNN:corresponding atomic positions - groundtruth_likelihoods_tables_SAC-CNN: likelihood maps generated by SAC-CNN in two different tabulated formats - groundtruth_labeled_images_all-models: images labeled with CV, TEMImagenet and SAC-CNN models (including Fe-GCN) - groundtruth_segmentation: segmentation of ground truth set for density calculation - all images (including ground truth set), as described above - all_original_images: false-color, png format - all_original_images_tif: as above, but in tif format (including pixel size) - all_labeled_images_SAC-CNN - all_atomic_positions_SAC-CNN - all_segmentation