Just Accepted
Just Accepted Articles have been posted online after technical editing and typesetting for immediate view. The final edited version with page numbers will appear in the Current Issue soon.
Submit a Manuscript
Deep learning-based workflow for atomic image denoising and chemical identification
Ke Ma, Shiqiang Feng, Haihui Hu, Yimeng Cai, Dechao Chen, Lili Han*
https://doi.org/10.1016/j.cjsc.2025.100572
ABSTRACT
Aberration-corrected annular dark-field scanning transmission electron microscopy (ADF-STEM) is a powerful tool for structural and chemical analysis of materials. Conventional analyses of ADF-STEM images rely on human labeling, making them labor-intensive and prone to subjective error. Here, we introduce a deep-learning-based workflow combining a pix2pix network for image denoising and either a mathematical algorithm local intensity threshold segmentation (LITS) or another deep learning network UNet for chemical identification. After the denoising, the processed images exhibit a five-fold improvement in signal-to-noise ratio and a 20% increase in accuracy of atomic localization. Then, we take atomic-resolution images of Y-Ce dual-atom catalysts (DACs) and Fe-doped ReSe2 nanosheets as examples to validate the performance. Pix2pix is applied to identify atomic sites in Y-Ce DACs with a location recall of 0.88 and a location precision of 0.99. LITS is used to further differentiate Y and Ce sites by the intensity of atomic sites. Furthermore, pix2pix and UNet workflow with better automaticity is applied to identification of Fe-doped ReSe2 nanosheets. Three types of atomic sites (Re, the substitution of Fe for Re, and the adatom of Fe on Re) are distinguished with the identification recall of more than 0.90 and the precision of more than 0.93. These results suggest that this strategy facilitates high-quality and automatically chemical identification of atomic-resolution images.
- Address: Fuzhou High & New Technology Industrial Zone, Shangjie, Minhou, Fuzhou, Fujian 350108, China
- Tel: (86)-591-63173769, 63173770
- E-mail: cjsc@fjirsm.ac.cn
CN 35–1112/TQ, ISSN 0254–5861 Copyright @ 2022 Chinese Journal of Structural Chemistry-www.Chinese Journal of Structural Chemistry.net. All Rights Reserved 闽ICP备2022002645号-1