Monte Carlo trained ANN framework for quantitative fluorescence correction in coupled absorbing-scattering media
Optics Express · 24.08.2026 · 34 (17) · 30956
DOI: 10.1364/oe.606305 ↗Kurz gefasst
Fachartikel in der Zeitschrift Optics Express, online erschienen am 10.08.2026. Verfasst von Yuxi Jiang und 8 weiteren Autor:innen. Die Arbeit wurde bisher 1-mal zitiert (Crossref) und nennt 48 Literaturangaben. Gefördert von: National Key Research and Development Program of China, Anhui Provincial Natural Science Foundation, Research Team Building Project of Hefei Comprehensive National Science Center.
Abstract (Originalsprache)
Fluorescence spectroscopy is highly sensitive for chemical analysis, but its quantitative accuracy deteriorates in turbid media because background absorption and multiple scattering distort photon transport and break the linear fluorescence-concentration relationship. Conventional correction methods generally treat absorption and scattering as separable effects, which limits their accuracy under strongly coupled optical conditions. To address this problem, we propose a Monte Carlo-trained ANN framework for quantitative fluorescence correction in coupled absorbing-scattering media. A customized Monte Carlo model is developed to simulate stochastic photon transport within the experimental geometry, incorporating focused excitation, wavelength-dependent absorption, multiple scattering, fluorescence generation, and finite-angle detection. Simulation data generated by the Monte Carlo model are then used to train a feed-forward artificial neural network, which serves as a fast surrogate of the Monte Carlo attenuation mapping. The network learns the nonlinear relationship between macroscopic optical coefficients and the fluorescence attenuation factor, enabling recovery of intrinsic analyte concentrations from distorted signals. Experimental results show that the Monte Carlo model accurately reproduces absorption-induced attenuation, scattering-dependent collection loss, and their coupled effects, with mean relative errors below 18% in low-to-moderate interference regimes. Compared with the standard stepwise analytical correction method, which degrades markedly as absorption and scattering increase and yields a recovery of only 0.3% in highly turbid and strongly absorbing media, the proposed MC-ANN framework achieves stable concentration recovery of 83.7%-110.5% in absorption-dominated conditions and retains the correct concentration order of magnitude in strongly coupled absorption-scattering regimes. This work provides a Monte Carlo-based surrogate correction strategy for quantitative fluorescence sensing in coupled absorbing-scattering media, where stochastic photon-transport modeling supplies the physical attenuation information and the ANN enables rapid prediction.
Förderung
- National Key Research and Development Program of China
- Anhui Provincial Natural Science Foundation
- Research Team Building Project of Hefei Comprehensive National Science Center
- Open Fund of National Engineering Research Center
Zitieren
APA
Jiang, Y., Yang, R., Zhao, N., Yin, G., Gao, M., Wang, Y., Sun, L., Wu, Z., & Liu, W. (2026). Monte Carlo trained ANN framework for quantitative fluorescence correction in coupled absorbing-scattering media. Optics Express, 34(17), 30956. https://doi.org/10.1364/oe.606305
Vancouver
Jiang Y, Yang R, Zhao N, Yin G, Gao M, Wang Y, et al. Monte Carlo trained ANN framework for quantitative fluorescence correction in coupled absorbing-scattering media. Optics Express. 2026;34(17):30956. doi:10.1364/oe.606305
BibTeX
@article{Jiang2026,
author = {Jiang, Yuxi and Yang, Ruifang and Zhao, Nanjing and Yin, Gaofang and Gao, Ming and Wang, Yingchong and Sun, Lan and Wu, Zongshi and Liu, Wenqing},
title = {Monte Carlo trained ANN framework for quantitative fluorescence correction in coupled absorbing-scattering media},
journal = {Optics Express},
year = {2026},
volume = {34},
number = {17},
pages = {30956},
doi = {10.1364/oe.606305}
}
EUVIMED