Computational efficiency of deep learning-based seismic fault interpretation considering context window and input resolution

Abstract

The application of deep learning to seismic fault interpretation is often constrained by computational costs. To address this, we propose a novel strategy that decouples the context window size (the spatial range of seismic observations) from the input resolution (the actual matrix dimensions fed into the model), systematically investigating their combined impact on computational efficiency and interpretation accuracy. Using field-acquired seismic data from two distinct coal mines, we trained a lightweight two-dimensional convolutional neural network (CNN) on samples extracted with varying context windows (8 × 8 to 64 × 64 pixels), which are then uniformly resized to a fixed low resolution of 8 × 8 pixels. Our results demonstrate that enlarging the context window consistently improved model performance, with the 64 × 64 window achieving the highest precision (99.46%) and fault continuity, even after downscaling. In contrast, a combined multi-scale training set did not outperform the best single-window model, indicating that effective multi-scale fusion requires more advanced architectural designs. Our workflow highlights that contextual information remains crucial for feature learning despite input standardization, and offers an efficient paradigm: large context window + small fixed input + lightweight network, that maintains high accuracy while significantly reducing computational costs. This approach provides a practical pathway for deploying deep learning models in resource-limited geophysical applications.

Keywords

Seismic fault interpretation / Deep learning / Convolutional neural network / Computational efficiency / Lightweight modeling

Funding

This work was supported by the National Key Research and Development Program of China (Grant number 2023YFB3211002) and the National Natural Science Foundation of China (Grant number 42274165).

Conflict of interest

Suping Peng is one of the Editor-in-Chiefs 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 they have no competing interests.

Author contributions

Conceptualization: Bowen Deng Data curation: Guangui Zou, Suping Peng Methodology: Bowen Deng Software: Bowen Deng Supervision: Suping Peng Writing–original draft: Bowen Deng Writing–review & editing: Guangui Zou, Suping Peng, Chengyang Han, Jingwen Xue

Availability of data

Seismic data, codes and neural networks associated with this research are available from the corresponding author upon reasonable request.

References

[1] Karpatne A , Ebert—Uphoff I , Ravela S , Babaie HA , Kumar V. Machine Learning for the geosciences: Challenges and opportunities. IEEE Trans Knowl Data Eng. 2019; 31(8): 1544-1554. doi: 10.1109/TKDE.2018.2861006

[2] Lary DJ , Alavi AH , Gandomi AH , Walker AL. Machine learning in geosciences and remote sensing. Geosci Front . 2016; 7(1): 3-10. doi: 10.1016/j.gsf.2015.07.003

[3] Yu S , Ma J. Deep learning for geophysics: Current and future trends. Rev Geophys . 2021; 59(3): e2021RG000742. doi: 10.1029/2021RG000742

[4] Di H , Shafiq MA , AlRegib G. Seismic—fault detection based on multiattribute support vector machine analysis. In: SEG Technical Program Expanded Abstracts 2017. Proceedings of the 87th SEG Annual International Meeting; September 24—29, 2017; Houston, Texas, USA. Society of Exploration Geophysicists; 2017: 2039-2044. doi: 10.1190/segam2017—17748277.1

[5] Ren K , Zou G , Zhang S , Peng S , Gong F , Liu Y. Fault identification and reliability evaluation using an SVM model based on 3—D seismic data volume. Geophys J Int. 2023; 234(1): 755-768. doi: 10.1093/gji/ggad095

[6] Han C , Zou G , Yeh HG , Gong F , Shi S , Chen H. Intelligent fault prediction with wavelet—SVM fusion in coal mine. Comput Geosci. 2025; 194: 105744. doi: 10.1016/j.cageo.2024.105744

[7] Ren K , Zou G , Peng S , Yeh HG , Deng B , Ji Y. Fault identification based on the kernel principal component analysis—genetic particle swarm optimization—support vector machine algorithm for seismic attributes in the Sihe Coal Mine, Qinshui Basin, China. Interpretation. 2022; 11(1): T59-T73. doi: 10.1190/int—2022—0039.1

[8] Li W , Yang C , Sun D. Mining geophysical parameters through decision—tree analysis to determine correlation with tropical cyclone development. Comput Geosci. 2009; 35(2): 309-316. doi: 10.1016/j.cageo.2008.02.032

[9] Li D , Peng S , Lu Y , Guo Y , Cui X. Seismic structure interpretation based on machine learning: A case study in coal mining. Interpretation. 2019; 7(3): SE69-SE79. doi: 10.1190/INT—2018—0208.1

[10] Ray A , Myer D. Bayesian geophysical inversion with transdimensional Gaussian process machine learning. Geophys J Int . 2019; 217(3): 1706-1726. doi: 10.1093/gji/ggz111

[11] Brown WM , Gedeon TD , Groves DI , Barnes RG. Artificial neural networks: A new method for mineral prospectivity mapping. Aust J Earth Sci. 2000; 47(4): 757-770. doi: 10.1046/j.1440—0952.2000.00807.x

[12] LeCun Y , Boser B , Denker JS , et al. Backpropagation Applied to Handwritten Zip Code Recognition. Neural Comput. 1989; 1(4): 541-551. doi: 10.1162/neco.1989.1.4.541

[13] Krizhevsky A , Sutskever I , Hinton GE. ImageNet classification with deep convolutional neural networks. Commun ACM. 2017; 60(6): 84-90. doi: 10.1145/3065386

[14] Bergen KJ , Johnson PA , de Hoop MV , Beroza GC. Machine learning for data—driven discovery in solid Earth geoscience. Science. 2019; 363(6433): eaau0323. doi: 10.1126/science.aau0323

[15] An Y , Du H , Ma S , et al. Current state and future directions for deep learning based automatic seismic fault interpretation: A systematic review. Earth Sci Rev. 2023; 243: 104509. doi: 10.1016/j.earscirev.2023.104509

[16] Waldeland AU , Jensen AC , Gelius LJ , Solberg AHS. Convolutional neural networks for automated seismic interpretation. Lead Edge. 2018; 37(7): 529-537. doi: 10.1190/tle37070529.1

[17] Ronneberger O , Fischer P , Brox T. U—Net: Convolutional networks for biomedical image segmentation. In: Navab N, Hornegger J, Wells W, Frangi A, eds. Medical Image Computing and Computer—Assisted Intervention—MICCAI 2015; Lecture Notes in Computer Science (9351). Proceedings of the MICCAI 2015, 18th International Conference; October 5—9, 2015; Munich, Germany. Springer; 2015: 234-241. doi: 10.1007/978—3—319—24574—4_28

[18] Huang J , Nowack RL. Machine learning using U—Net convolutional neural networks for the imaging of sparse seismic data. Pure Appl Geophys. 2020; 177(6): 2685-2700. doi: 10.1007/s00024—019—02412—z

[19] Peters B , Haber E , Granek J. Neural networks for geophysicists and their application to seismic data interpretation. Lead Edge . 2019; 38(7): 534-540. doi: 10.1190/tle38070534.1

[20] 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

[21] Cunha A , Pochet A , Lopes H , Gattass M. Seismic fault detection in real data using transfer learning from a convolutional neural network pre—trained with synthetic seismic data. Comput Geosci. 2020; 135: 104344. doi: 10.1016/j.cageo.2019.104344

[22] Deng B , Zou G , Peng S , She J , Han C , Liu Y. An approach of 2D convolutional neural network—based seismic data fault interpretation with linear annotation and pixel thinking. Geophys Prospect. 2024; 72(9): 3350-3370. doi: 10.1111/1365—2478.13606

[23] Li G , Li H , He S , Wang L. Multi—stage progressive network for seismic random noise suppression. J Seism Explor. 2025; 34(1): 43. doi: 10.36922/jse025240011

[24] Wang W , Chen H , Chang D , Wang X , Wang S , Li D. Dualbranch dense network for seismic background noise elimination. J Seism Explor. 2025; 34(5): 53. doi: 10.36922/jse025290038

[25] Parsania PS , Virparia PV. A comparative analysis of image interpolation algorithms. Int J Adv Res Comput Commun Eng. 2016; 5(1): 29-34. doi: 10.17148/IJARCCE.2016.5107

[26] Jakhetiya V , Kumar A , Tiwari AK. A survey on image interpolation methods. In: Jusoff K, Xie Y, eds. Volume 7546, Second International Conference On Digital Image Processing. Proceedings of the Second International Conference on Digital Image Processing; February 26—28, 2010; Singapore. Society of Photo—Optical Instrumentation Engineers (SPIE); 2010: 75461T. doi: 10.1117/12.855799

[27] Han D. Comparison of commonly used image interpolation methods. In: Proceedings of the 2nd International Conference on Computer Science and Electronics Engineering (ICCSEE 2013); Advances in Intelligent Systems Research Volume 34. 2nd International Conference on Computer Science and Electronics Engineering (ICCSEE 2013); March 22—23, 2013; Hangzhou, China. Atlantis Press; 2013: 1556-1559. doi: 10.2991/iccsee.2013.391

[28] Hu K , Zhang D , Xia M , Qian M , Chen B. LCDNet: Lightweighted cloud detection network for high—resolution remote sensing images. IEEE J Sel Top Appl Earth Obs Remote Sens. 2022; 15: 4809-4823. doi: 10.1109/JSTARS.2022.3181303

[29] Lu W , Chen SB , Shu QL , Tang J , Luo B. DecoupleNet: A lightweight backbone network with efficient feature decoupling for remote sensing visual tasks. IEEE Trans Geosci Remote Sens. 2024; 62: 4414613. doi: 10.1109/TGRS.2024.3465496

[30] Xiao Y , Xu T , Yu X , Fang Y , Li J. A lightweight fusion strategy with enhanced interlayer feature correlation for small object detection. IEEE Trans Geosci Remote Sens. 2024; 62: 1-11. doi: 10.1109/TGRS.2024.3457155