Education, Science, Technology, Innovation and Life
Open Access
Sign In
  1. Home
  2. Journals
  3. Journal of Electronics and Information Science
  4. Vol 11, Issue 1, 2026

Physics-Guided Self-Supervised Dual-Stream Transformer for Robust Optoelectronic Spectral Analytics

Download as PDF

DOI: 10.23977/jeis.2026.110102 | Downloads: 0 | Views: 8

Author(s)

Shaoyi Sun 1, Chunyu Ma 1

Affiliation(s)

1 School of Electronic and Optical Engineering, Nanjing University of Science and Technology, Nanjing, China

Corresponding Author

Shaoyi Sun

ABSTRACT

Optoelectronic sensing, such as reflectance or absorbance spectroscopy, enables non-contact measurement and provides informative signals for material characterization, quality inspection, and process monitoring. However, practical optoelectronic spectral analytics is often limited by scarce labels and distribution shifts caused by illumination variation, device drift, and measurement noise. To address these issues, a physics-guided self-supervised dual-stream Transformer framework is developed for robust learning from optoelectronic spectra. First, radiometric-consistent calibration is performed and physically meaningful spectral views are constructed, including calibrated spectra and derivative-enhanced representations. Second, a dual-stream Transformer encoder is designed to jointly model the complementary views, where cross-attention and gated fusion are adopted to adaptively aggregate spectral features. Third, self-supervised pretraining is introduced through masked spectral modeling and condition-invariant contrastive learning, enabling label-efficient representation learning. In addition, physics-regularized objectives, including illumination-invariance consistency and spectral smoothness priors, are incorporated during fine-tuning to improve generalization under cross-condition evaluation. Experimental results on optoelectronic spectral datasets demonstrate that the proposed method consistently improves predictive accuracy and robustness compared with representative baselines, particularly under low-label settings and cross-device testing.

KEYWORDS

Optoelectronic sensing; Spectral analytics; Self-supervised learning; Dual-stream Transformer; Physics-guided regularization; Domain robustness

CITE THIS PAPER

Shaoyi Sun, Chunyu Ma. Physics-Guided Self-Supervised Dual-Stream Transformer for Robust Optoelectronic Spectral Analytics. Journal of Electronics and Information Science (2026) Vol. 11: 10-20. DOI: http://dx.doi.org/10.23977/10.23977/jeis.2026.110102.

REFERENCES

[1] D. Passos, P. Mishra, "Perspectives on deep learning for near-infrared spectral data modelling," NIR News, vol. 33, no. 7–8, pp. 9–12, 2022. DOI: 10.1177/09603360221142821.
[2] A. Puleio, et al., "Calibration of spectra in presence of non-stationary background using unsupervised physics-informed deep learning," Scientific Reports, 2023. DOI: 10.1038/s41598-023-29371-9.
[3] A. P. Vorozhtsov, P. V. Kitina, "Contrastive representation learning for spectroscopy data analysis," Mendeleev Communications, 2024. DOI: 10.1016/j.mencom.2024.10.006.
[4] R. Bushuiev, A. Bushuiev, R. Samusevich, C. Brungs, et al., "Self-supervised learning of molecular representations from millions of tandem mass spectra using DreaMS,” Nature Biotechnology, 2025. DOI: 10.1038/s41587-025-02663-3.
[5] I. Bayrakli, E. Eken, "Compact laser spectroscopy-based sensor using a transformer-based model for analysis of multiple molecules," Applied Optics, vol. 63, no. 26, pp. 6941–6947, 2024. DOI: 10.1364/AO.534782.
[6] A. L. Bowler, M. P. Pound, N. J. Watson, "Domain Adaptation for In-Line Allergen Classification Using Near-Infrared Spectroscopy," Sensors, 2022. DOI: 10.3390/s22031084.
[7] F. B. Lavoie, G. Robert, A. Langlet, R. Gosselin, "Calibration transfer by likelihood maximization: A standard-free approach capable of handling non-overlapping wavelength ranges,” Chemometrics and Intelligent Laboratory Systems, 2023, vol. 234, p. 104766. DOI: 10.1016/j.chemolab.2023.104766.
[8] M. Alberts et al., "Leveraging infrared spectroscopy for automated structure elucidation," Communications Chemistry, 2024. DOI: 10.1038/s42004-024-01341-w.
[9] O. C. Koyun et al., “RamanFormer: A Transformer-Based Quantification Approach for Raman Mixture Components," ACS Omega, 2024. DOI: 10.1021/acsomega.3c06135.
[10] S. K. Paidi, P. Maheshwari, "RamanMAE: Masked Autoencoders Enable Efficient Molecular Imaging by Learning Biologically Meaningful Spectral Representations,” Analytical Chemistry, 2025. DOI: 10.1021/acs.analchem.5b01234.
[11] P. Rocabert-Oriols, C. Lo Conte, N. López, J. Heras-Domingo, "Multi-modal contrastive learning for chemical structure elucidation with VibraCLIP," Digital Discovery, 2025. DOI: 10.1039/d5dd00269a.
[12] S. Guglielmini, V. Banchieri, F. Scholkmann, M. Wolf, "Transformer-based deep learning model for predicting fNIRS short-channel signals," Neurophotonics, 2025. DOI: 10.1117/1.NPh.12.1.011003.
[13] C. Lange et al., "Comparing machine learning methods for chemical mixture analysis using Raman spectra from eight spectrometers," Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy, 2025. DOI: 10.1016/j.saa.2025.125861.
[14] C. Lange et al., "A Setup for Automatic Raman Measurements in High-Throughput Experimentation," Biotechnology and Bioengineering, 2025. DOI: 10.1002/bit.27283.
[15] N. Coca-Lopez et al., "Artificial Intelligence-Powered Raman Spectroscopy through Open Science and FAIR Principles," ACS Nano, 2025. DOI: 10.1021/acsnano.5b05625.

Subscription
E-Mail Alert
Most popular papers
Call for Special Issue
Publication Ethics & OA Statement
Plagiarism Policy
Contact Us
Downloads: 14102
Visits: 618047

Sponsors, Associates, and Links

All published work is licensed under a Creative Commons Attribution 4.0 International License.

Copyright © 2016 - 2031 Clausius Scientific Press Inc. All Rights Reserved.