AccScience Publishing / JSE / Online First / DOI: 10.36922/JSE026130057
Submit to JSE
Cite this article
12
Download
173
Views
Journal Browser
Volume | Year
Issue
Search
Forthcoming Issue
Current Issue
View All
News and Announcements
View All
ARTICLE

Self-supervised denoising of seismic common reflection point gathers with a learnable activation network

Chong Sun1,2 ,  Lang Yang1 ,  Yang Tan1 ,  Gui Chen2* ,  Yang Liu2 ,  Shikai Jian1,2 ,  Zilun Xiong1
Show Less
1 Tarim Oilfield, China National Petroleum Corporation, Korla, Xinjiang, China
2 State Key Laboratory of Petroleum Resources and Engineering, China University of Petroleum (Beijing), Changping, Beijing, China
JSE, 026130057 https://doi.org/10.36922/JSE026130057
Received: 27 March 2026 | Revised: 7 July 2026 | Accepted: 10 July 2026 | Published online: 23 July 2026
© 2026 by the Author(s). This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution 4.0 International License ( https://creativecommons.org/licenses/by/4.0/ )
Download PDF
XML
Cite
Abstract

Seismic common reflection point (CRP) gathers are often contaminated by noise, which degrades data quality and affects subsequent imaging and inversion. Supervised deep-learning denoising methods require noise-free labels, which are rarely available in field applications. We propose a patch-based self-supervised method for denoising seismic CRP gathers without a clean reference. The method uses a learnable activation denoising network, which combines fully connected layers, adaptive nonlinear activation functions, soft attention modules, and residual connections. The network is trained directly on overlapping noisy patches using the Softplus-smoothed log-cosh loss. Experiments on synthetic and field CRP gathers show that the proposed method suppresses noise while preserving coherent and weak reflection events. Compared to frequency-domain deconvolution, deep image prior, and Self2Self, the proposed method achieves a higher signal-to-noise ratio and lower root mean square error on synthetic data and shows weaker signal leakage in field-data evaluations. Additional tests show good performance for the tested Gaussian random noise, 0–80 Hz band-limited noise, and erratic-trace noise cases. However, the linear coherent-noise test shows only partial improvement, indicating that strongly coherent interference remains challenging. An amplitude-versus-angle comparison further suggests better preservation of the amplitude-versus-offset-related amplitude trend in the tested band-limited-noise case.

Keywords
Seismic common reflection point gathers
Random noise attenuation
Self-supervised deep learning
Learnable activations
Attention mechanism
Funding
This work was jointly supported by the Geophysical Exploration Technologies and Field Experiments for Ultra-Deep and Complex Oil and Gas Reservoirs under Grant No. YF202401 and the National Natural Science Foundation of China (NSFC) under Grant No. 42274147.
Conflict of interest
Yang Liu is an Editorial Board Member of this journal, but was not in any way involved in the editorial and peer-review process conducted for this paper, directly or indirectly. The authors declare that they have no competing interests.
References
  1. Liu Y, Liu C, Wang D. A 1D time-varying median filter for seismic random, spike-like noise elimination. Geophysics. 2009;74(1):V17-V24. doi: 10.1190/1.3043446
  2. Wang Y, Liu X, Gao F, Rao Y. Robust vector median filtering with a structure-adaptive implementation. Geophysics. 2020;85(5):V407-V414. doi: 10.1190/geo2020-0012.1
  3. Gülünay N. Signal leakage in f-x deconvolution algorithms. Geophysics. 2017;82(5):W31-W45. doi: 10.1190/geo2017-0007.1
  4. Liu G, Chen X. Noncausal f-x-y regularized nonstationary prediction filtering for random noise attenuation on 3D seismic data. J Appl Geophys. 2013;93:60-66. doi: 10.1016/j.jappgeo.2013.03.007
  5. Wang H, Chen W, Huang W, et al. Nonstationary predictive filtering for seismic random noise suppression: a tutorial. Geophysics. 2021;86(3):W21-W30. doi: 10.1190/geo2020-0368.1
  6. Alsdorf D. Noise reduction in seismic data using Fourier correlation coefficient filtering. Geophysics. 1997;62(5):1617-1627. doi: 10.1190/1.1444264
  7. Cao S, Chen X. The second-generation wavelet transform and its application in denoising of seismic data. Appl Geophys. 2005;2(2):70-74. doi: 10.1007/s11770-005-0034-4
  8. Anvari R, Siahsar MAN, Gholtashi S, Kahoo AR, Mohammadi M. Seismic random noise attenuation using synchrosqueezed wavelet transform and low-rank signal matrix approximation. IEEE Trans Geosci Remote Sens. 2017;55(11):6574-6581. doi: 10.1109/TGRS.2017.2730228
  9. Lian C, Chen W. Research on seismic noise attenuation method based on wavelet transform and fx-VMD. IEEE Access. 2025;13:33544-33554. doi: 10.1109/ACCESS.2025.3543607
  10. Shan H, Ma J, Yang H. Comparisons of wavelets, contourlets and curvelets in seismic denoising. J Appl Geophys. 2009;69(2):103-115. doi: 10.1016/j.jappgeo.2009.08.002
  11. Ma J, Plonka G. The curvelet transform. IEEE Signal Process Mag. 2010;27(2):118-133. doi: 10.1109/MSP.2009.935453
  12. Oropeza V, Sacchi M. Simultaneous seismic data denoising and reconstruction via multichannel singular spectrum analysis. Geophysics. 2011;76(3):V25-V32. doi: 10.1190/1.3552706
  13. Huang W, Wang R, Chen Y, Li H, Gan S. Damped multichannel singular spectrum analysis for 3D random noise attenuation. Geophysics. 2016;81(4):V261-V270. doi: 10.1190/geo2015-0264.1
  14. Li Z, Mo T, Song J, Wang B. Deblending and interpolation of subsampled blended seismic data based on damped randomized singular spectrum analysis. Geophys Prospect. 2024;72(6):2200-2213. doi: 10.1111/1365-2478.13507
  15. Beckouche S, Ma J. Simultaneous dictionary learning and denoising for seismic data. Geophysics. 2014;79(3):A27-A31. doi: 10.1190/geo2013-0382.1
  16. Liu D, Gao L, Wang X, Chen W. A dictionary learning method with atom splitting for seismic footprint suppression. Geophysics. 2021;86(6):V509-V523. doi: 10.1190/geo2020-0681.1
  17. Lan N, Zhang F, Sang K, Yin X. Simultaneous denoising and resolution enhancement of seismic data based on elastic convolution dictionary learning. Pet Sci. 2023;20(4):2127-2140. doi: 10.1016/j.petsci.2023.02.023
  18. Wu X, Liang L, Shi Y, Fomel S. FaultSeg3D: using synthetic data sets to train an end-to-end convolutional neural network for 3D seismic fault segmentation. Geophysics. 2019;84(3):IM35-IM45. doi: 10.1190/geo2018-0646.1
  19. Liu J, Cao J, Zhao L, You J, Li H. Super-resolution reconstruction of seismic images based on deep residual channel attention mechanism. IEEE Access. 2024;12:149032-149044. doi: 10.1109/ACCESS.2024.3477984
  20. Chen G, Liu Y, Zhang M, Sun Y, Zhang H. Unsupervised seismic reconstruction via deep learning with one-dimensional signal representation. Comput Geosci. 2025;200:105916. doi: 10.1016/j.cageo.2025.105916
  21. Saad OM, Alkhalifah T. F-SiameseFWI: a novel deep-learning framework for multisource full-waveform inversion. Geophysics. 2025;90(4):R221-R230. doi: 10.1190/geo2024-0785.1
  22. Liu Y, Sun Y, Zhang H, Tian W, Chen G, Ma J. Ren Gong Zhi Neng Di Zhen Zi Liao Chu Li Yu Jie Shi Fang Fa Yan Jiu Jin Zhan [A review of artificial intelligence-based seismic data processing and interpretation methods]. J Geophys Prospecting. 2025;1(1):1-23. [In Chinese] doi: 10.13810/j.cnki.issn.1000-7210.20240388
  23. Yu S, Ma J, Wang W. Deep learning for denoising. Geophysics. 2019;84(6):V333-V350. doi: 10.1190/geo2018-0668.1
  24. Yang L, Chen W, Liu W, Zha B, Zhu L. Random noise attenuation based on residual convolutional neural network in seismic datasets. IEEE Access. 2020;8:30271-30286. doi: 10.1109/ACCESS.2020.2972464
  25. Dong X, Li Y, Zhong T, Wu N, Wang H. Random and coherent noise suppression in DAS-VSP data by using a supervised deep learning method. IEEE Geosci Remote Sens Lett. 2022;19:1-5. doi: 10.1109/LGRS.2020.3023706
  26. Deng D, Liu Y, Chen G. Deep learning framework for the removal of low-frequency noise from reverse time migration. In: Proceedings of the International Meeting for Applied Geoscience & Energy. Society of Exploration Geophysicists; 2026:2478-2482. doi: 10.1190/image2025-4314502.1
  27. Zhang M, Liu Y, Bai M, Chen Y. Seismic noise attenuation using unsupervised sparse feature learning. IEEE Trans Geosci Remote Sens. 2019;57(12):9709-9723. doi: 10.1109/TGRS.2019.2928715
  28. Gao J, Li Z, Zhang M, Gao Y, Gao W. Unsupervised seismic random noise suppression based on local similarity and replacement strategy. IEEE Access. 2023;11:48924-48934. doi: 10.1109/ACCESS.2023.3272905
  29. Chen G, Liu Y. Unsupervised multi-dimensional seismic denoising and interpolation via single-trace deep representation learning. In: Proceedings of the International Meeting for Applied Geoscience & Energy. Society of Exploration Geophysicists; 2026:2373-2377. doi: 10.1190/image2025-4301721.1
  30. Liu N, Wang Z, Wang J, Wang D, Lou Y, Gao J. PD-VBS: Real seismic image denoising with pixel-shuffle down-sampling and visible blind-spots. IEEE Trans Geosci Remote Sens. 2024;62:1-10. doi: 10.1109/TGRS.2024.3401224
  31. Liu N, Wang J, Gao J, Chang S, Lou Y. Similarity-informed self-learning and its application on seismic image denoising. IEEE Trans Geosci Remote Sens. 2022;60:1-13. doi: 10.1109/TGRS.2022.3210217.
  32. Ulyanov D, Vedaldi A, Lempitsky V. Deep image prior. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. June 18-22, 2018; Salt Lake City, UT. IEEE Computer Society. 2018:9446-9454. doi: 10.1109/CVPR.2018.00984
  33. Qiu C, Wu B, Liu N, Zhu X, Ren H. Deep learning prior model for unsupervised seismic data random noise attenuation. IEEE Geosci Remote Sens Lett. 2022;19:1-5. doi: 10.1109/LGRS.2021.3053760
  34. Zhang Y, Wang B. Unsupervised seismic random noise attenuation by a recursive deep image prior. Geophysics. 2023;88(6):V473-V485. doi: 10.1190/geo2022-0612.1
  35. Oboué YASI, Chen Y, Guo Z, Chen Y. Leveraging overfitting for good—A two-step deep image prior model for seismic denoising. Geophysics. 2025;90(3):V205-V221. doi: 10.1190/geo2024-0236.1
  36. Li J, Trad D, Liu D. Robust seismic data denoising via self-supervised deep learning. Geophysics. 2024;89(5):V437-V451. doi: 10.1190/geo2023-0762.1
  37. Ozawa M. Enhancing seismic noise suppression using the Noise2Noise framework. Geophysics. 2025;90(2):V97-V110. doi: 10.1190/geo2024-0106.1
  38. Birnie C, Ravasi M, Liu S, Alkhalifah T. The potential of self-supervised networks for random noise suppression in seismic data. Artif Intell Geosci. 2021;2:47-59. doi: 10.1016/j.aiig.2021.11.001
  39. Fang W, Fu L, Li H, Liu S, Wang Q. BSnet: an unsupervised blind spot network for seismic data random noise attenuation. IEEE Trans Geosci Remote Sens. 2022;60:1-13. doi: 10.1109/TGRS.2022.3179718
  40. Chen G, Liu Y, Zhang M, Zhang H. Dropout-based robust self-supervised deep learning for seismic data denoising. IEEE Geosci Remote Sens Lett. 2022;19:1-5. doi: 10.1109/LGRS.2022.3167999
  41. Liu S, Birnie C, Alkhalifah T. Trace-wise coherent noise suppression via a self-supervised blind-trace deep-learning scheme. Geophysics. 2023;88(6):V459-V472. doi: 10.1190/geo2022-0371.1
  42. Zhang Y, Lin H, Li Y, Ma H. A patch based denoising method using deep convolutional neural network for seismic image. IEEE Access. 2019;7:156883-156894. doi: 10.1109/ACCESS.2019.2949774
  43. Saad OM, Chen Y. A fully unsupervised and highly generalized deep learning approach for random noise suppression. Geophys Prospect. 2021;69(4):709-726. doi: 10.1111/1365-2478.13062
  44. Saad OM, Ravasi M, Alkhalifah T. Self-supervised multi-stage deep learning network for seismic data denoising. Artif Intell Geosci. 2025;6(1):100123. doi: 10.1016/j.aiig.2025.100123
  45. Liu Z, Wang Y, Vaidya S, et al. KAN: Kolmogorov-Arnold Networks. arXiv. Preprint posted online 2025. doi: 10.48550/ARXIV.2404.19756
  46. Saleh RA, Saleh AKMdE. Statistical Properties of the log-cosh Loss Function Used in Machine Learning. arXiv. Preprint posted online 2024. doi: 10.48550/ARXIV.2208.04564
  47. Catania L, Allegra D. Redefining visual quality: the impact of loss functions on INR-based image compression. In: 2024 IEEE International Conference on Image Processing (ICIP). IEEE;2024:1973-1979. doi: 10.1109/ICIP51287.2024.10647328
  48. Martin GS, Wiley R, Marfurt KJ. Marmousi2: an elastic upgrade for Marmousi. Lead Edge. 2006;25(2):156-166. doi: 10.1190/1.2172306
  49. Fomel S. Local seismic attributes. Geophysics. 2007;72(3):A29-A33. doi: 10.1190/1.2437573
Share
Back to top
Journal of Seismic Exploration, Electronic ISSN: 0963-0651 Print ISSN: 0963-0651, Published by AccScience Publishing
We use cookies on our website to ensure you get the best experience.
Read more about our cookies here
Accept