Automatic baseline correction of 1D NMR spectra using a spline-constrained neural network
5 affiliations
- 1School of Electrical and Electronic Engineering, Wuhan Polytechnic University, Wuhan, 430023, China; State Key Laboratory of Magnetic Resonance Spectroscopy and Imaging, Innovation Academy for Precision Measurement Science and Technology, Chinese Academy of Sciences, Wuhan, 430071, China
- 2State Key Laboratory of Magnetic Resonance Spectroscopy and Imaging, Innovation Academy for Precision Measurement Science and Technology, Chinese Academy of Sciences, Wuhan, 430071, China
- 3School of Electrical and Electronic Engineering, Wuhan Polytechnic University, Wuhan, 430023, China; Corresponding author.
- 4School of Electrical and Electronic Engineering, Wuhan Polytechnic University, Wuhan, 430023, China
- 5State Key Laboratory of Magnetic Resonance Spectroscopy and Imaging, Innovation Academy for Precision Measurement Science and Technology, Chinese Academy of Sciences, Wuhan, 430071, China; Corresponding author.
Abstract
Baseline distortion remains a persistent source of error in NMR spectroscopy, compromising lineshape fidelity, weak-peak detection, and quantitative peak integration. Automatic baseline correction is particularly challenging for spectra with densely overlapped resonances, limited signal-free regions, or complex low-frequency distortions. Here, we propose a structurally model-constrained deep learning framework, named SplineNet, for automatic baseline correction of 1D NMR spectra. Instead of directly mapping a distorted spectrum to a corrected spectrum or an unconstrained point-wise baseline, the network predicts the coefficients of a cubic B-spline basis, from which the baseline is reconstructed. This output-space constraint incorporates the smooth and slowly varying nature of NMR baselines and reduces the tendency of the model to absorb narrow resonance peaks into the baseline estimate. The method was evaluated using simulated spectra with known baselines, water-suppressed 1 H NMR spectra, and 13 C DEPT-135 spectra, and was compared with conventional baseline-correction algorithms and unconstrained deep learning models. The spline-constrained model reduced baseline estimation errors, improved stability in peak-crowded regions, and yielded peak integrals closer to reference values than direct point-wise regression. These results suggest that constraining neural-network outputs to a spline-defined smooth function space provides a robust and physically interpretable strategy for automated NMR baseline correction.
Related
- Electronic excitation spectra and recovery of excited states with neural network wave functions
- Syntheses, X‐ray Crystal Structures, and Raman Spectra of Self‐Assembled XeO3 Chain and Network Complexes of Polytopic sp2‐ and sp3‐Nitrogen Bases
- Truncated automatic sparse differentiation for machine learning interatomic potentials
- Harnessing the VO2 Phase Transition for Automatic Gain Control in Transimpedance Amplifiers
- Nanotoxicology and Nanoparticle Safety in Biomedical Designs [Correction]
BibTeX
@article{wang2026automatic,
title = {{Automatic baseline correction of 1D NMR spectra using a spline-constrained neural network}},
author = {He Wang and Gang Chen and Chaixiang Liu and Zhen Huang and Wang Mei and Chaoyang Liu and Qingjia Bao},
journal = {Magnetic Resonance Letters},
year = {2026},
month = sep,
doi = {10.1016/j.mrl.2026.200308},
url = {https://doi.org/10.1016/j.mrl.2026.200308},
note = {CC BY-NC-ND},
}
RIS
TY - JOUR
TI - Automatic baseline correction of 1D NMR spectra using a spline-constrained neural network
AU - He Wang
AU - Gang Chen
AU - Chaixiang Liu
AU - Zhen Huang
AU - Wang Mei
AU - Chaoyang Liu
AU - Qingjia Bao
PY - 2026
DA - 2026/09/17
JO - Magnetic Resonance Letters
DO - 10.1016/j.mrl.2026.200308
UR - https://doi.org/10.1016/j.mrl.2026.200308
AB - Baseline distortion remains a persistent source of error in NMR spectroscopy, compromising lineshape fidelity, weak-peak detection, and quantitative peak integration. Automatic baseline correction is particularly challenging for spectra with densely overlapped resonances, limited signal-free regions, or complex low-frequency distortions. Here, we propose a structurally model-constrained deep learning framework, named SplineNet, for automatic baseline correction of 1D NMR spectra. Instead of directly mapping a distorted spectrum to a corrected spectrum or an unconstrained point-wise baseline, the network predicts the coefficients of a cubic B-spline basis, from which the baseline is reconstructed. This output-space constraint incorporates the smooth and slowly varying nature of NMR baselines and reduces the tendency of the model to absorb narrow resonance peaks into the baseline estimate. The method was evaluated using simulated spectra with known baselines, water-suppressed 1 H NMR spectra, and 13 C DEPT-135 spectra, and was compared with conventional baseline-correction algorithms and unconstrained deep learning models. The spline-constrained model reduced baseline estimation errors, improved stability in peak-crowded regions, and yielded peak integrals closer to reference values than direct point-wise regression. These results suggest that constraining neural-network outputs to a spline-defined smooth function space provides a robust and physically interpretable strategy for automated NMR baseline correction.
SN - 2772-5162
SN - 2097-0048
ER -