Simultaneous Denoising and Baseline Correction of Microplate Raman Spectra Using a Dual-Branch U-Net
- 1University of Ottawa , , , ,
- 2University of Waterloo , , , ,
Abstract
We present a dual-branch U-Net architecture for simultaneous baseline correction and denoising of Raman spectra. The network features a shared encoder that diverges into two specialized decoding heads for the Raman signal and for the baseline. The two heads are coupled with a cross-branch gating mechanism. The model can cross-confirm the peaks by comparing the recovered Raman signal with the baseline-corrected spectrum. Moreover, the model supports quantitative analysis by integrating the deep Raman-head reconstruction, Ideep, expressed in the detector’s intensity units. The model was trained entirely using a custom synthetic data engine explicitly designed to emulate automated HTS acquisitions from microplates. Comprehensive validation demonstrates robust peak recovery on synthetic spectra with signal-to-noise ratios (SNRs) as low as 5. The model recovers the Raman signal from highly noisy glycerol samples and from moderately noisy adenine sulfate samples. Quantitative analysis is conducted on guanine by integrating Ideep at a deep decoder layer. Transfer is tested on l-tyrosine, which is chemically unrelated to the training distribution.
Related
- Direct molecular formula prediction from tandem mass spectra with a multi-task-trained encoder and deterministic stoichiometric correction
- An integrated workflow for real-time chemoenzymatic CMP-sialic acid preparation and a bioorthogonal microplate sialyltransferase assay (MPSA)
- Standardized Microplate‐Based Assays for Quantifying Antioxidant Activity in Extracellular Vesicles and Nanovesicles Across Biological Kingdoms
- PANDA: Protein All-atom Nested-tree Denoising Architecture for End-to-End Generation
- Volumetric denoising enables high-throughput volume electron microscopy and efficient downstream analysis
BibTeX
@article{atia2026simultaneous,
title = {{Simultaneous Denoising and Baseline Correction of Microplate Raman Spectra Using a Dual-Branch U-Net}},
author = {Khaled S. R. Atia and Robert Hunter and Hanan Anis},
journal = {ACS Omega},
year = {2026},
month = oct,
doi = {10.1021/acsomega.6c04490},
url = {https://pubs.acs.org/acsodf/article/doi/10.1021/acsomega.6c04490/5455979/Simultaneous-Denoising-and-Baseline-Correction-of},
note = {CC BY-NC-ND 4.0},
}
RIS
TY - JOUR
TI - Simultaneous Denoising and Baseline Correction of Microplate Raman Spectra Using a Dual-Branch U-Net
AU - Khaled S. R. Atia
AU - Robert Hunter
AU - Hanan Anis
PY - 2026
DA - 2026/10/22
JO - ACS Omega
DO - 10.1021/acsomega.6c04490
UR - https://pubs.acs.org/acsodf/article/doi/10.1021/acsomega.6c04490/5455979/Simultaneous-Denoising-and-Baseline-Correction-of
AB - We present a dual-branch U-Net architecture for simultaneous baseline correction and denoising of Raman spectra. The network features a shared encoder that diverges into two specialized decoding heads for the Raman signal and for the baseline. The two heads are coupled with a cross-branch gating mechanism. The model can cross-confirm the peaks by comparing the recovered Raman signal with the baseline-corrected spectrum. Moreover, the model supports quantitative analysis by integrating the deep Raman-head reconstruction, Ideep, expressed in the detector’s intensity units. The model was trained entirely using a custom synthetic data engine explicitly designed to emulate automated HTS acquisitions from microplates. Comprehensive validation demonstrates robust peak recovery on synthetic spectra with signal-to-noise ratios (SNRs) as low as 5. The model recovers the Raman signal from highly noisy glycerol samples and from moderately noisy adenine sulfate samples. Quantitative analysis is conducted on guanine by integrating Ideep at a deep decoder layer. Transfer is tested on l-tyrosine, which is chemically unrelated to the training distribution.
SN - 2470-1343
ER -