EUVIMEDEUVIMED entdecken
Crossref

All-optical object classification using an edge-detecting spin-differential diffractive network

Yetao Shu, Laixi Sun, Yuhai Li, Tong Fu, Yubin Zhang, Rong Wu, Wenhua Gu, Li Li, Fang Wang, Xiaodong Yuan, Junwei Zhang

Optics Express · 09.02.2026 · 34 (3) · 3494

DOI: 10.1364/oe.579508 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Optics Express, online erschienen am 26.01.2026. Verfasst von Yetao Shu und 10 weiteren Autor:innen. Die Arbeit wurde bisher 4-mal zitiert (Crossref) und nennt 50 Literaturangaben. Gefördert von: National Natural Science Foundation of China, Natural Science Foundation of Jiangsu Province, Laser Fusion Research Center, China Academy of Engineering Physics.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

All-optical computing offers ultra-high speed, low-power consumption, and parallel processing capabilities, crucial for overcoming Moore's Law limitations. However, conventional single-wavelength diffractive deep neural networks (D 2 NN) face significant challenges in achieving synergistic optimization between high-precision optical edge-feature extraction and classification tasks. Here, an edge-detecting spin-differential diffractive neural network (ESD-DNN) is proposed for single-wavelength all-optical object classification. The network architecture is implemented through a Pancharatnam-Berry phase gradient metasurface to achieve rapid edge-feature extraction, while classification inference is accomplished by utilizing a spin-differential mechanism based on left-/right-handed circularly polarized (LCP/RCP) components. Through end-to-end optimization of the diffractive layers, the ESD-DNN achieves co-optimization of edge-feature extraction and classification, significantly improving accuracy while reducing computational costs. Numerical validations reveal that the single-layer ESD-DNN attains 97.5% (MNIST) and 87.5% (Fashion-MNIST) classification accuracy, surpassing traditional single-wavelength D 2 NN by 10.2% and 5.2%, respectively. Meanwhile, it achieves 5-fold higher computational efficiency while reducing time complexity by 80% compared to a four-layer D 2 NN. Remarkably, under extreme conditions such as moderate turbulence intensity or thermal lensing effects, the network maintains >90% classification accuracy (MNIST), demonstrating its exceptional environmental robustness. These findings pave the way for applications in artificial intelligence, satellite remote sensing, intelligent industrial inspection, and space optical communications.

Förderung

  • National Natural Science Foundation of China
  • Natural Science Foundation of Jiangsu Province
  • Laser Fusion Research Center, China Academy of Engineering Physics
  • Sichuan Provincial Science and Technology Support Program
  • Fundamental Research Funds for the Central Universities

Zitieren

APA

Shu, Y., Sun, L., Li, Y., Fu, T., Zhang, Y., Wu, R., Gu, W., Li, L., Wang, F., Yuan, X., & Zhang, J. (2026). All-optical object classification using an edge-detecting spin-differential diffractive network. Optics Express, 34(3), 3494. https://doi.org/10.1364/oe.579508

Vancouver

Shu Y, Sun L, Li Y, Fu T, Zhang Y, Wu R, et al. All-optical object classification using an edge-detecting spin-differential diffractive network. Optics Express. 2026;34(3):3494. doi:10.1364/oe.579508

BibTeX
@article{Shu2026,
  author = {Shu, Yetao and Sun, Laixi and Li, Yuhai and Fu, Tong and Zhang, Yubin and Wu, Rong and Gu, Wenhua and Li, Li and Wang, Fang and Yuan, Xiaodong and Zhang, Junwei},
  title = {All-optical object classification using an edge-detecting spin-differential diffractive network},
  journal = {Optics Express},
  year = {2026},
  volume = {34},
  number = {3},
  pages = {3494},
  doi = {10.1364/oe.579508}
}
Kostenlos starten →