基于轻量级深度神经网络的核磁共振波谱降噪

doi:10.3866/PKU.WHXB202310045

物理化学学报 >> 2025, Vol. 41 >> Issue (2): 100017.doi: 10.3866/PKU.WHXB202310045

基于轻量级深度神经网络的核磁共振波谱降噪

詹昊霖¹^,²^,*(), 房启元¹, 刘佳伟¹, 史晓琦², 陈心语¹, 黄玉清², 陈忠²

¹ 合肥工业大学仪器科学与光电工程学院生物医学工程系, 测量理论与精密仪器安徽省重点实验室, 合肥 230009
² 厦门大学电子科学系, 福建省等离子体与磁共振重点实验室, 固体表面物理化学国家重点实验室, 福建厦门 361005

收稿日期:2023-10-30 修回日期:2023-12-18 录用日期:2023-12-19 发布日期:2024-01-03
通讯作者: Email: hlzhan@hfut.edu.cn (詹昊霖)
基金资助:
国家自然科学基金(22204038)

Noise Reduction of Nuclear Magnetic Resonance Spectroscopy Using Lightweight Deep Neural Network

Haolin Zhan¹^,²^,*(), Qiyuan Fang¹, Jiawei Liu¹, Xiaoqi Shi², Xinyu Chen¹, Yuqing Huang², Zhong Chen²

¹ Department of Biomedical Engineering, Anhui Province Key Laboratory of Measuring Theory and Precision Instrument, School of Instrument Science and Opto-electronics Engineering, Hefei University of Technology, Hefei 230009, China
² Department of Electronic Science, Fujian Provincial Key Laboratory of Plasma and Magnetic Resonance, State Key Laboratory of Physical Chemistry of Solid Surfaces, Xiamen University, Xiamen 361005, Fujian Province, China

Received:2023-10-30 Revised:2023-12-18 Accepted:2023-12-19 Published:2024-01-03
Contact: Email: hlzhan@hfut.edu.cn (Haolin Zhan)
Supported by:
the National Natural Science Foundation of China(22204038)

1. WHXB202310045-SI-7p.pdf(638KB)

摘要/Abstract

摘要：

核磁共振(NMR)波谱是一种用于探测分子结构和提供定量分析的稳健的非侵入性表征技术。然而，进一步的NMR应用通常受到低灵敏度性能的限制，尤其是对于异核实验。在此，我们提出了一种轻量级的深度学习协议，用于高质量、可靠和快速的NMR波谱降噪。该深度学习(DL)协议具有轻量级的网络优势和快速的计算效率，有效地抑制噪声和伪峰信号，并恢复几乎完全淹没在严重噪声中的目标弱峰，从而实现了可观的信噪比提升。此外，它仅使用物理驱动的仿真NMR数据学习，在频域中实现令人满意的波谱去噪，并允许区分真实信号和噪声伪影。此外，训练的轻量级网络模型通用于一维和多维NMR波谱，并适用于不同的化学样品。因此，本研究呈现的深度学习方法在化学、生物学、材料和生命科学等领域具有应用潜力。

关键词: 核磁共振波谱, 人工智能, 深度学习, 谱图去噪, 轻量级网络

Abstract:

Nuclear magnetic resonance (NMR) spectroscopy serves as a robust non-invasive characterization technique for probing molecular structure and providing quantitative analysis, however, further NMR applications are generally confined by the low sensitivity performance, especially for heteronuclear experiments. Herein, we present a lightweight deep learning protocol for high-quality, reliable, and very fast noise reduction of NMR spectroscopy. Along with the lightweight network advantages and fast computational efficiency, this deep learning (DL) protocol effectively reduces noises and spurious signals, and recovers desired weak peaks almost entirely drown in severe noise, thus implementing considerable signal-to-noise ratio (SNR) improvement. Additionally, it enables the satisfactory spectral denoising in the frequency domain and allows one to distinguish real signals and noise artifacts using solely physics-driven synthetic NMR data learning. Besides, the trained lightweight network model is general for one-dimensional and multi-dimensional NMR spectroscopy, and can be exploited on diverse chemical samples. As a result, the deep learning method presented in this study holds potential applications in the fields of chemistry, biology, materials, life sciences, and among others.

Key words: NMR spectroscopy, Artificial intelligence, Deep learning, Spectral denoising, Lightweight network

詹昊霖, 房启元, 刘佳伟, 史晓琦, 陈心语, 黄玉清, 陈忠. 基于轻量级深度神经网络的核磁共振波谱降噪[J]. 物理化学学报, 2025, 41(2), 100017. doi: 10.3866/PKU.WHXB202310045

Haolin Zhan, Qiyuan Fang, Jiawei Liu, Xiaoqi Shi, Xinyu Chen, Yuqing Huang, Zhong Chen. Noise Reduction of Nuclear Magnetic Resonance Spectroscopy Using Lightweight Deep Neural Network[J]. Acta Phys. -Chim. Sin. 2025, 41(2), 100017. doi: 10.3866/PKU.WHXB202310045

链接本文: https://www.whxb.pku.edu.cn/CN/10.3866/PKU.WHXB202310045

https://www.whxb.pku.edu.cn/CN/Y2025/V41/I2/100017

图/表 6

图1

图2

表1

图3

图4

图5

参考文献 49

1	Theillet F. X. Chem. Rev. 2022, 122(10), 9497. doi: 10.1021/acs.chemrev.1c00937
2	Chen K.; Zornes A.; Nguyen V.; Wang B.; Gan Z. H.; Crossley S. P.; White J. L. J. Am. Chem. Soc. 2022, 144(37), 16916. doi: 10.1021/jacs.2c05332
3	Xin J. X.; Wei D. X.; Ren Y.; Wang J. L.; Yang G.; Zhang H.; Li J.; Fu C.; Yao Y. F. Magn. Reson. Med. 2022, 89(5), 1728. doi: 10.1002/mrm.29562
4	Zhan H. L.; Ji L. F.; Cao S. H.; Feng Y.; Jiang Y. X.; Huang Y. Q.; Sun S. G.; Chen Z. Chin. J. Catal. 2023, 53, 171. doi: 10.1016/S1872-2067(23)64526-7
5	Zhan H. L.; Gao C. Y.; Huang C. D.; Lin X. Q.; Huang Y. Q.; Chen Z. Anal. Chim. Acta 2023, 1277, 341682. doi: 10.1016/j.aca.2023.341682
6	Zhan H. L.; Hao M. Y.; Feng Y.; Cao S. H.; Ni Z. K.; Huang Y. Q.; Chen Z. J. Phys. Chem. Lett. 2021, 12(3), 1073. doi: 10.1021/acs.jpclett.0c03549
7	Xu J.; Deng F. Acta Phys. -Chim. Sin. 2020, 36(4), 1912074. doi: 10.3866/PKU.WHXB201912074
	徐君; 邓风. 物理化学学报, 2020, 36(4), 1912074. doi: 10.3866/PKU.WHXB201912074
8	Hu R.; Wei L. Y.; Xian J. L.; Fang G. Y.; Wu Z. A.; Fan M.; Guo J. Y.; Li Q. X.; Liu K. S.; Jiang H. Y.; et al Acta Phys. -Chim. Sin. 2023, 39(9), 2212025. doi: 10.3866/PKU.WHXB202212025
	胡荣; 韦丽云; 鲜靖林; 房光钰; 吴植傲; 樊淼; 郭家越; 李青翔; 刘凯思; 姜会钰; 等. 物理化学学报, 2023, 39(9), 2212025. doi: 10.3866/PKU.WHXB202212025
9	He Z. J.; Huang M.; Lin T. J.; Zhong L. S. Acta Phys. -Chim. Sin. 2023, 39(9), 2212060. doi: 10.3866/PKU.WHXB202212060
	何展军; 黄敏; 林铁军; 钟良枢. 物理化学学报, 2023, 39(9), 2212060. doi: 10.3866/PKU.WHXB202212060
10	Gan Z. H.; Hung I.; Wang X. L.; Paulino J.; Wu G.; Litvak I. M.; Gor'kov P. L.; Brey W. W.; Lendi P.; Schiano J. L.; et al J. Magn. Reson. 2017, 284, 125. doi: 10.1016/j.jmr.2017.08.007
11	Chen K. Z.; Horstmeier S.; Nguyen V. T.; Wang B.; Crossley S. P.; Pham T.; Gan Z. H.; Hung I.; White J. L. J. Am. Chem. Soc. 2020, 142(16), 7514. doi: 10.1021/jacs.0c00590
12	Kovacs H.; Moskau D.; Spraul M. Prog. Nucl. Magn. Reson. Spectrosc. 2005, 46(2–3), 131. doi: 10.1016/j.pnmrs.2005.03.001
13	Zhang R. C.; Mroue K. H.; Ramamoorthy A. J. Magn. Reson. 2016, 266, 59. doi: 10.1016/j.jmr.2016.03.006
14	Zhou Y.; van Zijl P. C. M.; Xu X.; Xu J. D.; Li Y. G.; Chen L.; Yadav N. N. Proc. Natl. Acad. Sci. U. S. A. 2020, 117(6), 3144. doi: 10.1073/pnas.1909921117
15	Sonnefeld A.; Razanahoera A.; Pelupessy P.; Bodenhausen G.; Sheberstov K. Sci. Adv. 2022, 8, eade2113. doi: 10.1126/sciadv.ade2113
16	Pang Z. F.; Xi G. H.; Gao L.; Cao W. C.; Yin J. L.; Kong X. Q. Acta Phys. -Chim. Sin. 2020, 36(4), 1906018. doi: 10.3866/PKU.WHXB201906018
	庞振峰; 管晗曦; 高李娜; 曹伟成; 尹竟琳; 孔学谦. 物理化学学报, 2020, 36(4), 1906018. doi: 10.3866/PKU.WHXB201906018
17	Elliott S. J.; Stern Q.; Ceillier M.; El Darai T.; Cousin S. F.; Cala O.; Jannin S. Prog. Nucl. Magn. Reson. Spectrosc. 2021, 126–127, 59. doi: 10.1016/j.pnmrs.2021.04.002
18	Kharbanda Y.; Urbańczyk M.; Zhivonitko V. V.; Mailhiot S.; Kettunen M. I.; Telkki V.-V. Angew. Chem. Int. Ed. 2022, 61(28), e202203957. doi: 10.1002/anie.202203957
19	Jaroszewicz M. J.; Liu M.; Kim J.; Zhang G.; Kim Y.; Hilty C.; Frydman L. Nat. Commun. 2022, 13(1), 833. doi: 10.1038/s41467-022-28304-w
20	Szekely O.; Olsen G. L.; Novakovic M.; Rosenzweig R.; Frydman L. J. Am. Chem. Soc. 2020, 142(20), 9267. doi: 10.1021/jacs.0c00807
21	Marshall H.; Stewart N. J.; Chan H. F.; Rao M.; Norquay G.; Wild J. M. Prog. Nucl. Magn. Reson. Spectrosc. 2021, 122, 42. doi: 10.1016/j.pnmrs.2020.11.002
22	Li H. D.; Zhao X. C.; Wang Y. J.; Lou X.; Chen S. Z.; Deng H.; Shi L.; Xie J. S.; Tang D. Z.; Zhao J. P.; et al Sci. Adv. 2021, 7(1), eabc8180. doi: 10.1126/sciadv.abc8180
23	Green R. A.; Adams R. W.; Duckett S. B.; Mewis R. E.; Williamson D. C.; Green G. G. Prog. Nucl. Magn. Reson. Spectrosc. 2012, 67, 1. doi: 10.1016/j.pnmrs.2012.03.001
24	Eills J.; Cavallari E.; Carrera C.; Budker D.; Aime S.; Reineri F. J. Am. Chem. Soc. 2019, 141(51), 20209. doi: 10.1021/jacs.9b10094
25	Barskiy D. A.; Knecht S.; Yurkovskaya A. V.; Ivanov K. L. Prog. Nucl. Magn. Reson. Spectrosc. 2019, 114–115, 33. doi: 10.1016/j.pnmrs.2019.05.005
26	Koprivica D.; Martinho R. P.; Novakovic M.; Jaroszewicz M. J.; Frydman L. J. Magn. Reson. 2022, 338, 107187. doi: 10.1016/j.jmr.2022.107187
27	Qiu T. Y.; Liao W. J.; Huang Y. H.; Wu J. Y.; Guo D.; Liu D. B.; Wang X.; Cai J.-F.; Hu B. W.; Qu X. B. IEEE Trans. Instrum. Meas. 2021, 70, 1. doi: 10.1109/tim.2021.3109743
28	Jiang B.; Luo F.; Ding Y. M.; Sun P.; Zhang X.; Jiang L. G.; Li C.; Mao X. A.; Yang D. W.; Tang C.; et al Anal. Chem. 2013, 85(4), 2523. doi: 10.1021/ac303726p
29	Kusaka Y.; Hasegawa T.; Kaji H. J. Phys. Chem. A 2019, 123(47), 10333. doi: 10.1021/acs.jpca.9b04437
30	Froeling M.; Prompers J. J.; Klomp D. W. J.; van der Velden T. A. Magn. Reson. Med. 2021, 85(6), 2992. doi: 10.1002/mrm.28654
31	LeCun Y.; Bengio Y.; Hinton G. Nature 2015, 521(7553), 436. doi: 10.1038/nature14539
32	Manu V. S.; Olivieri C.; Veglia G. Nat. Commun. 2023, 14(1), 4144. doi: 10.1038/s41467-023-39581-4
33	Wang, W. L.; Ma, L. H.; Maletic-Savatic, M.; Liu, Z. D. NMRQNet: a deep learning approach for automatic identification and quantification of metabolites using Nuclear Magnetic Resonance (NMR) in human plasma samples. bioRxiv [Preprint], 2023. Available Online: https://www.ncbi.nlm.nih.gov/pubmed/36909516 (accessed on Mar 2, 2023).
34	Qu X. B.; Huang Y. H.; Lu H. F.; Qiu T. Y.; Guo D.; Agback T.; Orekhov V.; Chen Z. Angew. Chem. Int. Ed. 2020, 59(26), 10297. doi: 10.1002/anie.201908162
35	Zheng X. X.; Yang Z. X.; Yang C.; Shi X. Q.; Luo Y.; Luo J.; Zeng Q.; Lin Y. Q.; Chen Z. J. Phys. Chem. Lett. 2022, 13(9), 2101. doi: 10.1021/acs.jpclett.2c00100
36	Karunanithy G.; Hansen D. F. J. Biomol. NMR 2021, 75(4–5), 179. doi: 10.1007/s10858-021-00366-w
37	Karunanithy G.; Mackenzie H. W.; Hansen D. F. J. Am. Chem. Soc. 2021, 143(41), 16935. doi: 10.1021/jacs.1c04010
38	Chen B.; Wu L. B.; Cui X. H.; Lin E. P.; Cao S. H.; Zhan H. L.; Huang Y. Q.; Yang Y.; Chen Z. Anal. Chem. 2023, 95(31), 11596. doi: 10.1021/acs.analchem.3c00537
39	Lee H. H.; Kim H. Magn. Reson. Med. 2019, 82(1), 33. doi: 10.1002/mrm.27727
40	Chen D. C.; Hu W. Q.; Liu H. T.; Zhou Y. R.; Qiu T. Y.; Huang Y. H.; Wang Z.; Lin M. J.; Lin L. J.; Wu Z. G.; et al IEEE T. Comput. Imag. 2023, 9, 448. doi: 10.1109/tci.2023.3267623
41	Wu K.; Luo J.; Zeng Q.; Dong X.; Chen J. Y.; Zhan C. Q.; Chen Z.; Lin Y. Q. Anal. Chem. 2021, 93(3), 1377. doi: 10.1021/acs.analchem.0c03087
42	Ronneberger, O.; Fischer, P.; Brox, T. U-Net: Convolutional networks for biomedical image segmentation. In Medical Image Computing and Computer-Assisted Intervention – MICCAI 2015, 18th International Conference, Munich, Germany, Oct. 5–9, 2015; Navab, N., Hornegger, J., Wells, W. M., Frangi, A. F., Eds.; Springer Nature: Berlin, Germany, 2015; pp. 234–241.
43	Stoller, D.; Ewert, S.; Dixon, S. A multi-scale neural network for end-to-end audio source separation. arxiv [Preprint], 2018. Available Online: https://arxiv.org/abs/1806.03185 (accessed on Jun 8, 2018).
44	Macartney, C.; Weyde, T. Improved speech enhancement with the Wave-U-Net. arXiv[Preprint], 2018. Available Online: https://arxiv.org/abs/1811.11307 (accessed on Nov 27, 2018).
45	Gao J.; Liang E.; Ma R. S.; Li F. D.; Liu Y. X.; Liu J.; Jiang L.; Li C. G.; Dai H. M.; Wu J. H.; et al Angew. Chem. Int. Ed. 2017, 56(42), 12982. doi: 10.1002/anie.201707114
46	Rethage, D.; Pons, J.; Serra, X. A Wavenet for speech denoising. In ICASSP 2018–2018 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), Calgary, AB, Canada, Apr. 15–20, 2018; IEEE: New York, US, 2018; pp. 5069–5073.
47	Zangger K. Prog. Nucl. Magn. Reson. Spectrosc. 2015, 86(86–87), 1. doi: 10.1016/j.pnmrs.2015.02.002
48	Zhan H. L.; Huang Y. Q.; Chen Z. J. Phys. Chem. Lett. 2019, 10(23), 7356. doi: 10.1021/acs.jpclett.9b03092
49	Zhan H. L.; Hao M. Y.; Lin E. P.; Zheng Z. Y.; Huang C. D.; Cai S. H.; Cao S. H.; Huang Y. Q.; Chen Z. Anal. Chem. 2023, 95(2), 1002. doi: 10.1021/acs.analchem.2c03678

Methods	Training times (h)	Model sizes (MB)	Parameters (MB)	Reconstruction times (ms)	SNR enhancements	RMSDs (e-4)	R² (%)	Error rates ^b
LD-Net (L = 3)	5.98	0.82	0.20	32.3 ± 2.50	139.4 ± 46.47	9.2 ± 2.47	99.85 ± 0.13	0%
LD-Net (L = 12)	4.32	38.76	10.13	46.1 ± 1.84	140.7 ± 16.05	8.4 ± 2.27	99.86 ± 0.13	0.5%
DN-Unet ⁴¹	11.90	568.30	148.93	225.2 ± 12.11	238.1 ± 38.56	8.4 ± 2.28	99.85 ± 0.12	0.5%
Wavenet ⁴⁶	46.34	8.63	2.24	130.1 ± 5.36	97.1±40.87	9.3 ± 2.43	99.86 ± 0.12	0.5%
U-Net ⁴²	1.55	56.50	14.79	33.3 ± 1.01	135.1±47.28	11.2 ± 2.92	99.86 ± 0.11	14%
CoSeM ²⁶	–			3.3e5 ± 1.58e4	3.2 ± 0.70 ^c	63.9 ± 0.042	99.51±0.27	0%

[1]	张恒睿, 徐熹骏, 李璕琭, 杲祥文. 生成式人工智能在电池研究中的应用：现状与展望[J]. 物理化学学报, 2025, 41(10): 100115 - .
[2]	薛智敏,严传玉,赵新辉,于东琨,牟天成. Hofmeister离子对水溶液中热响应聚合物的局域环境的影响[J]. 物理化学学报, 2019, 35(1): 49 -57 .
[3]	林梅钦;宋锦宏;唐亚林;董朝霞;李明远;吴肇亮. 柠檬酸铝溶液中铝的形态研究[J]. 物理化学学报, 2005, 21(02): 202 -204 .

基于轻量级深度神经网络的核磁共振波谱降噪

Noise Reduction of Nuclear Magnetic Resonance Spectroscopy Using Lightweight Deep Neural Network

RichHTML

PDF

赞

补充材料

可视化

摘要/Abstract

引用本文

使用本文

图/表 6

参考文献 49

相关文章 3

Metrics

推荐阅读 0