Transformer neural network-based algorithm for manufacturing error inversion of optical coatings
A Transformer-based deep learning algorithm is proposed to efficiently inverse thickness errors in optical coatings, achieving spectral differences under 1% with computations in milliseconds after training on simulated datasets.
Authors
- Xinshang Niu (ORCID: https://orcid.org/0000-0001-6326-2954)
- Zhanshan Wang (ORCID: https://orcid.org/0000-0002-6022-4933)
- Hongfei Jiao (ORCID: https://orcid.org/0000-0003-2331-9420)
- Xiaochuan Ji
- Xinbin Cheng (ORCID: https://orcid.org/0000-0002-3855-483X)
- Jianglin Dai
- Jinlong Zhang (ORCID: https://orcid.org/0000-0002-1334-6436)
Institutions
- Tongji University (CN)
Publication Details
- Published
- 2025-01-01
- DOI
- https://doi.org/10.1364/oic.2025.me.10
- Primary Topic
- Surface Roughness and Optical Measurements
- Type
- article
- Field-Weighted Citation Impact
- 0.00