Deep learning for three-dimensional seismic fault prediction using convolutional neural networks

Abstract

Understanding seismic faults is essential for generating prospects, modeling reservoirs, and assessing carbon dioxide storage. Identifying faults in complex tectonic regimes presents significant challenges, especially in areas that have undergone multiple phases of tectonic activity. Even with advances in structural seismic attributes and machine learning, interpreters often rely on manual methods to examine complex fault systems. This work introduces a method for predicting three-dimensional seismic faults using convolutional neural networks (CNNs), effectively overcoming the constraints of conventional interpretation techniques. The research uses CNNs to demonstrate the effectiveness of seismic attributes in training models that identify faults with high accuracy and consistency. This approach, unlike manual interpretation, reduces time, costs, and subjective errors by leveraging automated learning techniques, thereby improving reproducibility and efficiency while reducing interpreter bias. The research highlights the growing importance of solid computational tools in geophysics, especially as seismic datasets become more complex and wider. The approach significantly enhances confidence in artificial intelligence-assisted geological analysis by validating its performance with real-world data. The validation accuracy rises from 0.936 to 0.9436 across configurations, while the validation loss increases from 0.5635 to 0.9941 across diverse patches of the trained model.

Keywords

Deep learning / Convolutional neural networks / Fault likelihood / Fault cube / Fault prediction

References

[1] Khosro Anjom F, Vaccarino F, Socco LV. Machine learning for seismic exploration: Where are we and how far are we from the holy grail? Geophysics. 2024; 89(1): WA157-WA178. doi: 10.1190/geo2023—0129.1

[2] Xiong W, Ji X, Ma Y, et al. Seismic fault detection with convolutional neural network. Geophysics . 2018; 83(5): O97-O103. doi: 10.1190/geo2017—0666.1

[3] Babikir I, Elsaadany M. Machine learning—based seismic characterization of deepwater turbidites in the Dangerous Grounds area, Northwest Sabah, offshore Malaysia. Acta Geophys. 2024; 73(1): 379-391. doi: 10.1007/s11600—024—01396—2

[4] Bashir Y, Khan M, Mahgoub M, et al. Machine Learning Application on Seismic Diffraction Detection and Preservation for High Resolution Imaging. In: International Petroleum Technology Conference. IPTC; 2024:IPTC—23668. doi: 10.2523/IPTC—23668—EA

[5] Ismail A, Radwan AA, Leila M, Abdelmaksoud A, Ali M. Unsupervised machine learning and multi—seismic attributes for fault and fracture network interpretation in the Kerry Field, Taranaki Basin, New Zealand. Geomech Geophys Geo—Energy Geo—Resour. 2023; 9(1). doi: 10.1007/s40948—023—00646—9

[6] Khan M, Bery AA, Bashir Y, Sharoni SMH, Ali SS. Automated Fault Network Extraction in Complex Tectonic Regimes: A Hybrid Machine Learning and Structural Attributes Approach. Appl Comput Geosci. 2025; 27: 100264. doi: 10.1016/j.acags.2025.100264

[7] Wu X, Shi Y, Fomel S, Liang L. Convolutional neural networks for fault interpretation in seismic images. In: SEG Technical Program Expanded Abstracts 2018. Tulsa, OK: Society of Exploration Geophysicists; 2018: 1946-1950. doi: 10.1190/segam2018—2995341.1

[8] Guo Y, Peng S, Du W, Li D. Fault and horizon automatic interpretation by CNN: A case study of coalfield. J Geophys Eng. 2020; 17(6): 1016-1025. doi: 10.1093/jge/gxaa060

[9] Owusu BA, Boateng CD, Asare VDS, Danuor SK, Adenutsi CD, Quaye JA. Seismic facies analysis using machine learning techniques: a review and case study. Earth Sci Inform . 2024; 17(5): 3899-3924. doi: 10.1007/s12145—024—01395—3

[10] Zhao T, Jayaram V, Roy A, Marfurt KJ. A comparison of classification techniques for seismic facies recognition. Interpretation . 2015; 3(4): SAE29-SAE58. doi: 10.1190/INT—2015—0044.1

[11] Imran QS, Siddiqui NA, Latiff AHA, et al. Automated fault detection and extraction under gas chimneys using hybrid discontinuity attributes. Appl Sci. 2021; 11(16): 7218. doi: 10.3390/app11167218

[12] Khan M, Bery AA, Bashir Y, Sharoni SMH, Gnapragasan J, Imran QS. Optimizing petrophysical property prediction in fluvial—deltaic reservoirs: a multi—seismic attribute transformation and probabilistic neural network approach. J Pet Explor Prod Technol. 2025; 15(2): 29. doi: 10.1007/s13202—024—01912—6

[13] Kamsari M, Sadeghi S, de Oliveira GG, et al. The role of data augmentation and attention mechanisms in UNet and ConvNeXt architectures for optimizing breast tumor segmentation. Sci Rep. 2025; 15: 45268. doi: 10.1038/s41598—025—29841—2

[14] Anari S, Ranjbarzadeh R, Cunneen M, Bendechache M. Privacy—Preserving Federated Learning for Human Intention Modeling in Pediatric Cerebral Palsy Using Extended Reality. In: 2025 IEEE 49th Annual Computers, Software, and Applications Conference (COMPSAC). IEEE; 2025: 1-6. doi: 10.1109/COMPSAC65507.2025.00209

[15] Ranjbarzadeh R, Anari S, Cunneen M, Bendechache M. Lightweight Deep Learning with Virtual Reality Visualization for Offline Tumor Segmentation in Rural Environments. In: 2025 IEEE 49th Annual Computers, Software, and Applications Conference (COMPSAC). IEEE; 2025: 1577-1582. doi: 10.1109/COMPSAC65507.2025.00211

[16] Bagherian Kasgari A, Sadeghi S, Zarbakhsh P, Jafarzadeh Ghoushchi S, Ranjbarzadeh R. A spiking convolutional neural network for glioma brain tumor segmentation using a spike—timing—dependent plasticity method. Neurocomputing . 2026; 660: 131903. doi: 10.1016/j.neucom.2025.131903

[17] Zeng A, Yan L, Huang Y, Ren E, Liu T, Zhang H. Intelligent detection of small faults using a support vector machine. Energies. 2021; 14(19): 6242. doi: 10.3390/en14196242

[18] Bashir Y, Babasafari AA, Alashloo SYM, Muztaza NM, Ali SH, Imran QS. Seismic wave propagation characteristics using conventional and advance modelling algorithm for d—data imaging. J Seism Explor. 2021; 30: 21-44.

[19] Zhao T, Chen G, Pang C, et al. Hybrid convolutional neural network—graph attention network—gradient boosting decision tree model for seismic impedance inversion prediction. J Seism Explor. 2025; 34(5): 81-98. doi: 10.36922/JSE025310051

[20] Pang C, Zhao T, Chen G, et al. Earthquake and blast recognition based on CEEMDAN multiscale fuzzy entropy and NSGAIII optimized 1D—CNN neural networks. J Seism Explor. 2025; 34(1): 22-42. doi: 10.36922/JSE025260029

[21] An Y, Guo J, Ye Q, Childs C, Walsh J, Dong R. Deep convolutional neural network for automatic fault recognition from 3D seismic datasets. Comput Geosci . 2021; 153: 104776. doi: 10.1016/j.cageo.2021.104776

[22] Ter Borgh MM, Jaarsma B, Rosendaal EA. Structural development of the northern Dutch offshore: Paleozoic to present. In: Geological Society, London, Special Publications. Vol 471. London, UK: Geological Society; 2019: 115-131. doi: 10.1144/SP471.4

[23] Lima G, Zeiser FA, Da Silveira A, Rigo S, de Oliveira Ramos G. An encoder—decoder deep neural network for binary segmentation of seismic facies. Comput Geosci. 2024; 183: 105507. doi: 10.1016/j.cageo.2023.105507

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

[25] Wei XL, Zhang CX, Kim SW, et al. Seismic fault detection using convolutional neural networks with focal loss. Comput Geosci. 2022; 158: 104968. doi: 10.1016/j.cageo.2021.104968

[26] Jozinović D, Lomax A, Štajduhar I, Michelini A. Rapid prediction of earthquake ground shaking intensity using raw waveform data and a convolutional neural network. Geophys J Int. 2020; 222(2): 1379-1389. doi: 10.1093/gji/ggaa233