EUVIMEDEUVIMED entdecken
Crossref

Optical model compression learning for galaxy morphology classification with diffractive deep neural networks

Zongyu Lu, Jinye Li, Mingxuan Li, Shiyu Mei, Songyuan He, Jing Zhang, You Yu, Buwen Cheng, Jianguo Liu

Optics Express · 04.05.2026 · 34 (9) · 16168

DOI: 10.1364/oe.587209 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Optics Express, online erschienen am 22.04.2026. Verfasst von Zongyu Lu und 8 weiteren Autor:innen. Die Arbeit nennt 25 Literaturangaben. Gefördert von: National Natural Science Foundation of China.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Galaxy morphology is key to understanding cosmic evolution. While large astronomical surveys such as the Sloan Digital Sky Survey produce vast amounts of data, energy–efficient processing remains a challenge. Diffractive deep neural networks (D 2 NNs) offer a promising optical computing solution with high throughput and low power consumption. However, deep D 2 NNs suffer from optical attenuation, while shallow networks have limited representation capability. To address this trade–off, we extend the concept of knowledge distillation with optical–domain considerations and develop an optical model compression learning (OMCL) framework for D 2 NNs. Numerical simulations on SDSS galaxy images show that the compressed model achieves 69.25% classification accuracy, representing a 27.5% improvement over conventional training, and exhibits partial recognition ability for unseen categories. Further analysis reveals that the compressed model develops smoother and sparser phase patterns, leading to improved robustness to noise and stronger tolerance to low–precision phase quantization (e.g., 3–bit). These results indicate that transferring optical–domain features from deeper networks can effectively mitigate the limitations of shallow D 2 NNs, providing a practical pathway toward fabrication–friendly and high–performance optical neural networks for large–scale astronomical data analysis.

Förderung

  • National Natural Science Foundation of China

Zitieren

APA

Lu, Z., Li, J., Li, M., Mei, S., He, S., Zhang, J., Yu, Y., Cheng, B., & Liu, J. (2026). Optical model compression learning for galaxy morphology classification with diffractive deep neural networks. Optics Express, 34(9), 16168. https://doi.org/10.1364/oe.587209

Vancouver

Lu Z, Li J, Li M, Mei S, He S, Zhang J, et al. Optical model compression learning for galaxy morphology classification with diffractive deep neural networks. Optics Express. 2026;34(9):16168. doi:10.1364/oe.587209

BibTeX
@article{Lu2026,
  author = {Lu, Zongyu and Li, Jinye and Li, Mingxuan and Mei, Shiyu and He, Songyuan and Zhang, Jing and Yu, You and Cheng, Buwen and Liu, Jianguo},
  title = {Optical model compression learning for galaxy morphology classification with diffractive deep neural networks},
  journal = {Optics Express},
  year = {2026},
  volume = {34},
  number = {9},
  pages = {16168},
  doi = {10.1364/oe.587209}
}
Kostenlos starten →