Beyond the thin-layer limit: differentiable volumetric training for visible-range diffractive neural networks
Optics Express · 16.11.2026 · 34 (23) · 43118
DOI: 10.1364/oe.609758 ↗Kurz gefasst
Fachartikel in der Zeitschrift Optics Express, online erschienen am 02.11.2026. Verfasst von Dineth Jayakody und 1 weiteren Autor:innen. Die Arbeit nennt 22 Literaturangaben. Gefördert von: National Institute of General Medical Sciences.
Abstract (Originalsprache)
Diffractive deep neural networks ( D 2 NNs) could provide compact, power-efficient optical front-ends for machine vision, but visible-range systems remain less experimentally mature than terahertz demonstrations. A key modeling challenge is the thin-layer approximation commonly used during training, which neglects field evolution within finite-thickness relief structures. We introduce a differentiable beam-propagation (∂BPM) layer that represents each diffractive element as a finite-thickness volume while preserving end-to-end trainability without full-wave simulation in the training loop. After accounting for axial-reference-plane alignment and simple global calibration effects, the thin and volumetric descriptions converge for sufficiently thin reliefs, while task-dependent residual mismatch remains as thickness increases, especially for imaging and lower-index structures. Volumetric fine-tuning reduces this residual, and same-structure FDTD comparisons show that ∂BPM follows the full-wave field more closely than the thin-layer approximation. These results establish ∂BPM as a more faithful differentiable forward model for finite-thickness D 2 NN design while identifying regimes where the thin approximation remains adequate.
Förderung
- National Institute of General Medical Sciences
Zitieren
APA
Jayakody, D., & Wadduwage, D. N. (2026). Beyond the thin-layer limit: differentiable volumetric training for visible-range diffractive neural networks. Optics Express, 34(23), 43118. https://doi.org/10.1364/oe.609758
Vancouver
Jayakody D, Wadduwage DN. Beyond the thin-layer limit: differentiable volumetric training for visible-range diffractive neural networks. Optics Express. 2026;34(23):43118. doi:10.1364/oe.609758
BibTeX
@article{Jayakody2026,
author = {Jayakody, Dineth and Wadduwage, Dushan N.},
title = {Beyond the thin-layer limit: differentiable volumetric training for visible-range diffractive neural networks},
journal = {Optics Express},
year = {2026},
volume = {34},
number = {23},
pages = {43118},
doi = {10.1364/oe.609758}
}
EUVIMED