{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,4,1]],"date-time":"2026-04-01T22:46:22Z","timestamp":1775083582866,"version":"3.50.1"},"reference-count":53,"publisher":"Society of Exploration Geophysicists","issue":"1","funder":[{"DOI":"10.13039\/501100001809","name":"National Natural Science Foundation of China","doi-asserted-by":"publisher","award":["U23B20157"],"award-info":[{"award-number":["U23B20157"]}],"id":[{"id":"10.13039\/501100001809","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/501100012226","name":"Fundamental Research Funds for the Central Universities","doi-asserted-by":"publisher","id":[{"id":"10.13039\/501100012226","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":["library.seg.org"],"crossmark-restriction":true},"short-container-title":[],"published-print":{"date-parts":[[2026,1,1]]},"abstract":"<jats:title>ABSTRACT<\/jats:title>\n                  <jats:p>Reliable lithofacies prediction from seismic data was essential for advancing exploration and characterization in hydrocarbon reservoirs, CO2 storage, geothermal energy recovery, and groundwater management. Lithofacies prediction from seismic data under a data-driven supervised deep learning framework was often subject to issues of geological prior inconsistency, poor generalizability, and weak interpretability. To address these challenges, this study proposed the multi-seismic with multi-geological constraint Net (MSMGNet), a deep learning-based lithofacies prediction approach that not only incorporated multi-seismic information but also took into account prior geological knowledge, including stratigraphic differences, prior lithologic probabilities, lithofacies transition probabilities, and thickness distributions of the lithofacies. The input-based strategy and loss-based strategy were proposed to incorporate statistical geological features for constraining lithofacies prediction. In the input-based approach, geological features were embedded as structured input vectors, where stratigraphic units were encoded using one-hot encoding, and lithofacies proportions, transition probabilities, and thickness distributions were vectorized into numerical feature representations. In the loss-based approach, these geological features were formulated as regularization terms within the loss function, where Kullback\u2013Leibler (KL) divergence was used to penalize deviations between predicted and prior distributions. Cross-well blind tests in a complex coal-bearing clastic reservoir demonstrated that incorporating these statistical geological features significantly improved lithofacies prediction performance. More importantly, geology-constrained deep learning approaches enhanced the ability to capture lithological variations between stratigraphic units, characterize thickness distributions, and identify thin sandstone and coal layers. The input-based MSMGNet, which integrated all four geological constraints, achieved the highest prediction accuracy and geological consistency in comparison with the baseline model without geological constraints and the loss-based MSMGNet.<\/jats:p>","DOI":"10.1190\/geo2024-0800","type":"journal-article","created":{"date-parts":[[2025,9,25]],"date-time":"2025-09-25T21:47:52Z","timestamp":1758836872000},"page":"MR33-MR50","update-policy":"https:\/\/doi.org\/10.1190\/crossmark-policy","source":"Crossref","is-referenced-by-count":0,"title":["Seismic lithofacies prediction via deep learning with prior geological constraints"],"prefix":"10.1190","volume":"91","author":[{"ORCID":"https:\/\/orcid.org\/0000-0001-5581-2493","authenticated-orcid":false,"given":"Luanxiao","family":"Zhao","sequence":"first","affiliation":[{"name":"Tongji University 1 , , Shanghai , .","place":["China"]},{"name":"State Key Laboratory of Marine Geology 1 , , Shanghai , .","place":["China"]}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-3044-4281","authenticated-orcid":false,"given":"Jinyu","family":"Meng","sequence":"additional","affiliation":[{"name":"Tongji University 1 , , Shanghai , .","place":["China"]},{"name":"State Key Laboratory of Marine Geology 1 , , Shanghai , .","place":["China"]},{"name":"Tongji University 2 , , Shanghai , .","place":["China"]},{"name":"School of Ocean and Earth Science 2 , , Shanghai , .","place":["China"]}]},{"given":"Minghui","family":"Xu","sequence":"additional","affiliation":[{"name":"Tongji University 1 , , Shanghai , .","place":["China"]},{"name":"State Key Laboratory of Marine Geology 1 , , Shanghai , .","place":["China"]},{"name":"Tongji University 2 , , Shanghai , .","place":["China"]},{"name":"School of Ocean and Earth Science 2 , , Shanghai , .","place":["China"]}]},{"given":"Wenji","family":"Wang","sequence":"additional","affiliation":[{"name":"Shanghai Branch of CNOOC China Limited 3 , Shanghai , .","place":["China"]}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-0411-1192","authenticated-orcid":false,"given":"Jianhua","family":"Geng","sequence":"additional","affiliation":[{"name":"Tongji University 1 , , Shanghai , .","place":["China"]},{"name":"State Key Laboratory of Marine Geology 1 , , Shanghai , .","place":["China"]},{"name":"Tongji University 2 , , Shanghai , .","place":["China"]},{"name":"School of Ocean and Earth Science 2 , , Shanghai , .","place":["China"]}]}],"member":"186","published-online":{"date-parts":[[2026,2,24]]},"reference":[{"key":"2026040117175504700_R1","first-page":"2298","article-title":"Semi-supervised learning for acoustic impedance inversion, SEG Technical Program Expanded Abstracts 2019","author":"Alfarraj","year":"2019","journal-title":"Society of Exploration Geophysicists"},{"issue":"2","key":"2026040117175504700_R2","doi-asserted-by":"crossref","first-page":"C291","DOI":"10.1137\/23M1574051","article-title":"Neural DAEs: Constrained neural networks","volume":"47","author":"Boesen","year":"2025","journal-title":"SIAM Journal on Scientific Computing"},{"issue":"5","key":"2026040117175504700_R3","doi-asserted-by":"crossref","first-page":"75A165","DOI":"10.1190\/1.3478209","article-title":"Seismic inversion for reservoir properties combining statistical rock physics and geostatistics: A review","volume":"75","author":"Bosch","year":"2010","journal-title":"Geophysics"},{"issue":"3","key":"2026040117175504700_R4","doi-asserted-by":"crossref","first-page":"C13","DOI":"10.1190\/1.2842150","article-title":"Bayesian lithology and fluid prediction from seismic prestack data","volume":"73","author":"Buland","year":"2008","journal-title":"Geophysics"},{"key":"2026040117175504700_R5","doi-asserted-by":"crossref","first-page":"109836","DOI":"10.1016\/j.petrol.2021.109836","article-title":"Multi-parameter pre-stack seismic inversion based on deep learning with sparse reflection coefficient constraints","volume":"209","author":"Cao","year":"2022","journal-title":"Journal of Petroleum Science and Engineering"},{"key":"2026040117175504700_R6","article-title":"Principles of sequence stratigraphy","author":"Catuneanu","year":"2022","journal-title":"Newnes"},{"issue":"4","key":"2026040117175504700_R7","doi-asserted-by":"crossref","first-page":"WA213","DOI":"10.1190\/geo2019-0261.1","article-title":"Deep learning reservoir porosity prediction based on multilayer long short-term memory network","volume":"85","author":"Chen","year":"2020","journal-title":"Geophysics"},{"issue":"8","key":"2026040117175504700_R8","doi-asserted-by":"crossref","first-page":"5932","DOI":"10.1109\/TGRS.2020.2973171","article-title":"Physics-constrained deep learning of geomechanical logs","volume":"58","author":"Chen","year":"2020","journal-title":"IEEE Transactions on Geoscience and Remote Sensing"},{"issue":"3","key":"2026040117175504700_R9","doi-asserted-by":"crossref","first-page":"IM63","DOI":"10.1190\/geo2020-0449.1","article-title":"Imposing interpretational constraints on a seismic interpretation convolutional neural network","volume":"86","author":"Di","year":"2021","journal-title":"Geophysics"},{"issue":"3","key":"2026040117175504700_R10","doi-asserted-by":"crossref","first-page":"379","DOI":"10.1023\/B:MATG.0000028443.75501.d9","article-title":"Estimation of geological attributes from a well log: An application of hidden Markov chains","volume":"36","author":"Eidsvik","year":"2004","journal-title":"Mathematical Geology"},{"issue":"5","key":"2026040117175504700_R11","doi-asserted-by":"crossref","first-page":"569","DOI":"10.1023\/A:1011044812133","article-title":"A Markov chain model for subsurface characterization: Theory and applications","volume":"33","author":"Elfeki","year":"2001","journal-title":"Mathematical Geology"},{"issue":"4","key":"2026040117175504700_R12","doi-asserted-by":"crossref","first-page":"M185","DOI":"10.1190\/geo2022-0363.1","article-title":"Seismic predictions of fluids via supervised deep learning: Incorporating various class-rebalance strategies","volume":"88","author":"Gao","year":"2023","journal-title":"Geophysics"},{"key":"2026040117175504700_R13","doi-asserted-by":"crossref","DOI":"10.1002\/9781119086215","volume-title":"Seismic reservoir modeling: Theory, examples, and algorithms","author":"Grana","year":"2021"},{"issue":"4","key":"2026040117175504700_R14","doi-asserted-by":"crossref","first-page":"M53","DOI":"10.1190\/geo2011-0417.1","article-title":"Stochastic inversion of facies from seismic data based on sequential simulations and probability perturbation method","volume":"77","author":"Grana","year":"2012","journal-title":"Geophysics"},{"key":"2026040117175504700_R15","doi-asserted-by":"crossref","first-page":"220","DOI":"10.1190\/1.1444899","article-title":"Use of multiattribute transforms to predict log properties from seismic data","volume":"66","author":"Hampson","year":"2001","journal-title":"Geophysics"},{"issue":"8","key":"2026040117175504700_R16","doi-asserted-by":"crossref","first-page":"1735","DOI":"10.1162\/neco.1997.9.8.1735","article-title":"Long short-term memory","volume":"9","author":"Hochreiter","year":"1997","journal-title":"Neural Computation"},{"issue":"1","key":"2026040117175504700_R17","doi-asserted-by":"crossref","first-page":"147","DOI":"10.1142\/S0219530519410148","article-title":"A regularized convolutional neural network for semantic image segmentation","volume":"19","author":"Jia","year":"2021","journal-title":"Analysis and Applications"},{"issue":"1","key":"2026040117175504700_R18","doi-asserted-by":"crossref","first-page":"D31","DOI":"10.1190\/geo2022-0770.1","article-title":"Lithofacies identification from well-logging curves via integrating prior knowledge into deep learning","volume":"89","author":"Jiang","year":"2024","journal-title":"Geophysics"},{"issue":"5","key":"2026040117175504700_R19","doi-asserted-by":"crossref","first-page":"IM85","DOI":"10.1190\/geo2020-0676.1","article-title":"Lithology identification from well-log curves via neural networks with additional geologic constraint","volume":"86","author":"Jiang","year":"2021","journal-title":"Geophysics"},{"key":"2026040117175504700_R20","doi-asserted-by":"crossref","first-page":"88","DOI":"10.1016\/j.media.2019.02.009","article-title":"Constrained-CNN losses for weakly supervised segmentation","volume":"54","author":"Kervadec","year":"2019","journal-title":"Medical Image Analysis"},{"key":"2026040117175504700_R21","article-title":"Adam: A method for stochastic optimization","author":"Kingma","year":"2014","journal-title":"arXiv Preprint arXiv"},{"key":"2026040117175504700_R22","first-page":"25","volume-title":"Advances in neural information processing systems","author":"Krizhevsky","year":"2012"},{"issue":"1","key":"2026040117175504700_R23","doi-asserted-by":"crossref","first-page":"79","DOI":"10.1007\/BF02047072","article-title":"Markov chains and embedded Markov chains in geology","volume":"1","author":"Krumbein","year":"1969","journal-title":"Journal of the International Association for Mathematical Geology"},{"issue":"7553","key":"2026040117175504700_R24","doi-asserted-by":"crossref","first-page":"436","DOI":"10.1038\/nature14539","article-title":"Deep learning","volume":"521","author":"LeCun","year":"2015","journal-title":"Nature"},{"issue":"4","key":"2026040117175504700_R25","doi-asserted-by":"crossref","first-page":"W1","DOI":"10.1190\/geo2015-0572.1","article-title":"Tutorial: Spectral bandwidth extension \u2014 Invention versus harmonic extrapolation","volume":"82","author":"Liang","year":"2017","journal-title":"Geophysics"},{"issue":"1","key":"2026040117175504700_R26","doi-asserted-by":"crossref","first-page":"WA81","DOI":"10.1190\/geo2022-0148.1","article-title":"Unsupervised contrastive learning for seismic facies characterization","volume":"88","author":"Li","year":"2023","journal-title":"Geophysics"},{"issue":"1","key":"2026040117175504700_R27","doi-asserted-by":"crossref","first-page":"101311","DOI":"10.1016\/j.gsf.2021.101311","article-title":"Integrating deep learning and logging data analytics for lithofacies classification and 3D modeling of tight sandstone reservoirs","volume":"13","author":"Liu","year":"2022","journal-title":"Geoscience Frontiers"},{"key":"2026040117175504700_R28","first-page":"1","article-title":"Unsupervised deep learning for ground roll and scattered noise attenuation","volume":"61","author":"Liu","year":"2023","journal-title":"IEEE Transactions on Geoscience and Remote Sensing"},{"issue":"3","key":"2026040117175504700_R29","article-title":"Joint inversion of geophysical data for geologic carbon sequestration monitoring: A differentiable physics-informed neural network model","volume":"128","author":"Liu","year":"2023","journal-title":"Journal of Geophysical Research"},{"issue":"7","key":"2026040117175504700_R30","article-title":"Geostatistical inversion for subsurface characterization using Stein variational gradient descent with autoencoder neural network: An application to geologic carbon sequestration","volume":"129","author":"Liu","year":"2024","journal-title":"Journal of Geophysical Research"},{"issue":"4","key":"2026040117175504700_R31","doi-asserted-by":"crossref","first-page":"2991","DOI":"10.1007\/s12145-023-01079-4","article-title":"A comprehensive review of seismic inversion based on neural networks","volume":"16","author":"Li","year":"2023","journal-title":"Earth Science Informatics"},{"key":"2026040117175504700_R32","first-page":"552","article-title":"Extension of the limit theorems of probability theory to a sum of variables connected in a chain","volume":"1","author":"Markov","year":"1971","journal-title":"Dynam Probabilist Syst"},{"key":"2026040117175504700_R33","article-title":"Imposing hard constraints on deep networks: Promises and limitations","author":"M\u00e1rquez-Neila","year":"2017","journal-title":"arXiv Preprint arXiv"},{"issue":"1","key":"2026040117175504700_R34","doi-asserted-by":"crossref","first-page":"107","DOI":"10.1111\/1365-2478.13229","article-title":"Seismic impedance inversion using a multi-input neural network with a two-step training strategy","volume":"72","author":"Meng","year":"2024","journal-title":"Geophysical Prospecting"},{"key":"2026040117175504700_R35","article-title":"A primal dual formulation for deep learning with constraints","volume":"32","author":"Nandwani","year":"2019","journal-title":"Advances in Neural Information Processing Systems"},{"key":"2026040117175504700_R36","doi-asserted-by":"crossref","article-title":"Constrained convolutional neural networks for weakly supervised segmentation","author":"Pathak","DOI":"10.1109\/ICCV.2015.209"},{"key":"2026040117175504700_R37","first-page":"2","article-title":"Point-to-set distance functions for output-constrained neural networks","volume":"4","author":"Peters","year":"2022","journal-title":"Journal of Applied & Numerical Optimization"},{"key":"2026040117175504700_R38","first-page":"1","article-title":"Reservoir lithology identification based on improved adversarial learning","volume":"20","author":"Song","year":"2023","journal-title":"IEEE Geoscience and Remote Sensing Letters"},{"key":"2026040117175504700_R39","doi-asserted-by":"crossref","first-page":"129493","DOI":"10.1016\/j.jhydrol.2023.129493","article-title":"GANSim-surrogate: An integrated framework for stochastic conditional geomodelling","volume":"620","author":"Song","year":"2023","journal-title":"Journal of Hydrology"},{"key":"2026040117175504700_R40","first-page":"1","article-title":"LsmGANs: Image-domain least-squares migration using a new framework of generative adversarial networks","volume":"61","author":"Sun","year":"2023","journal-title":"IEEE Transactions on Geoscience and Remote Sensing"},{"issue":"2","key":"2026040117175504700_R41","doi-asserted-by":"crossref","first-page":"R21","DOI":"10.1190\/1.3294570","article-title":"Improved resolution in Bayesian lithology\/fluid inversion from prestack seismic data and well observations: Part 1\u2014Methodology","volume":"75","author":"Ulvmoen","year":"2010","journal-title":"Geophysics"},{"issue":"23","key":"2026040117175504700_R42","doi-asserted-by":"crossref","DOI":"10.1073\/pnas.2219573120","article-title":"Sensing prior constraints in deep neural networks for solving exploration geophysical problems","volume":"120","author":"Wu","year":"2023","journal-title":"Proceedings of the National Academy of Sciences"},{"issue":"5","key":"2026040117175504700_R43","doi-asserted-by":"crossref","first-page":"M151","DOI":"10.1190\/geo2021-0554.1","article-title":"Joint use of multiseismic information for lithofacies prediction via supervised convolutional neural networks","volume":"87","author":"Xu","year":"2022","journal-title":"Geophysics"},{"issue":"6","key":"2026040117175504700_R44","doi-asserted-by":"crossref","first-page":"1919","DOI":"10.1007\/s10712-023-09779-8","article-title":"Deep learning with fully convolutional and dense connection framework for ground roll attenuation","volume":"44","author":"Yang","year":"2023","journal-title":"Surveys in Geophysics"},{"key":"2026040117175504700_R45","first-page":"1","article-title":"A multi-task learning method for relative geologic time, horizons, and faults with prior information and transformer","volume":"61","author":"Yang","year":"2023","journal-title":"IEEE Transactions on Geoscience and Remote Sensing"},{"issue":"3","key":"2026040117175504700_R46","doi-asserted-by":"crossref","DOI":"10.1029\/2021RG000742","article-title":"Deep learning for geophysics: Current and future trends","volume":"59","author":"Yu","year":"2021","journal-title":"Reviews of Geophysics"},{"key":"2026040117175504700_R47","doi-asserted-by":"crossref","first-page":"219","DOI":"10.1016\/B978-0-12-822295-9.00009-1","volume-title":"Advances in Subsurface Data Analytics: Elsevier","author":"Zhang","year":"2022"},{"issue":"1","key":"2026040117175504700_R48","doi-asserted-by":"crossref","first-page":"13","DOI":"10.1109\/LGRS.2020.2964611","article-title":"Seismic lithology\/fluid prediction via a hybrid ISD-CNN","volume":"18","author":"Zhang","year":"2021","journal-title":"IEEE Geoscience and Remote Sensing Letters"},{"issue":"7","key":"2026040117175504700_R49","doi-asserted-by":"crossref","first-page":"5526","DOI":"10.1109\/TGRS.2020.3022368","article-title":"A spatially coupled data-driven approach for lithology\/fluid prediction","volume":"59","author":"Zhang","year":"2021","journal-title":"IEEE Transactions on Geoscience and Remote Sensing"},{"key":"2026040117175504700_R50","first-page":"1368","article-title":"Deep learning for seismic lithology prediction","volume":"215","author":"Zhang","year":"2018","journal-title":"Geophysical Journal International"},{"issue":"5","key":"2026040117175504700_R51","doi-asserted-by":"crossref","first-page":"M25","DOI":"10.1190\/geo2013-0406.1","article-title":"Probabilistic lithofacies prediction from prestack seismic data in a heterogeneous carbonate reservoir","volume":"79","author":"Zhao","year":"2014","journal-title":"Geophysics"},{"issue":"4","key":"2026040117175504700_R52","doi-asserted-by":"crossref","first-page":"827","DOI":"10.1111\/1365-2478.12043","article-title":"Quantitative geophysical pore-type characterization and its geological implication in carbonate reservoirs","volume":"61","author":"Zhao","year":"2013","journal-title":"Geophysical Prospecting"},{"issue":"4","key":"2026040117175504700_R53","doi-asserted-by":"crossref","first-page":"M151","DOI":"10.1190\/geo2020-0521.1","article-title":"Fluid and lithofacies prediction based on integration of well-log data and seismic inversion: A machine-learning approach","volume":"86","author":"Zhao","year":"2021","journal-title":"Geophysics"}],"container-title":["Geophysics"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/pubs.geoscienceworld.org\/seg\/geophysics\/article-pdf\/91\/1\/MR33\/7455586\/geo-2024-0800.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"syndication"},{"URL":"https:\/\/pubs.geoscienceworld.org\/seg\/geophysics\/article-pdf\/91\/1\/MR33\/7455586\/geo-2024-0800.pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,4,1]],"date-time":"2026-04-01T21:18:10Z","timestamp":1775078290000},"score":1,"resource":{"primary":{"URL":"https:\/\/pubs.geoscienceworld.org\/geophysics\/article\/91\/1\/MR33\/670100\/Seismic-lithofacies-prediction-via-deep-learning"},"secondary":[{"URL":"https:\/\/pubs.geoscienceworld.org\/geophysics\/article\/doi\/10.1190\/geo2024-0800\/670100\/seismic-lithofacies-prediction-via-deep-learning","label":"geoscienceworld"}]},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,1,1]]},"references-count":53,"journal-issue":{"issue":"1","published-print":{"date-parts":[[2026,1,1]]}},"URL":"https:\/\/doi.org\/10.1190\/geo2024-0800","relation":{},"ISSN":["0016-8033","1942-2156"],"issn-type":[{"value":"0016-8033","type":"print"},{"value":"1942-2156","type":"electronic"}],"subject":[],"published-other":{"date-parts":[[2026,1]]},"published":{"date-parts":[[2026,1,1]]},"assertion":[{"value":"2024-11-03","order":0,"name":"received","label":"Received","group":{"name":"publication_history","label":"Publication History"}},{"value":"2025-08-11","order":2,"name":"accepted","label":"Accepted","group":{"name":"publication_history","label":"Publication History"}},{"value":"2025-09-20","order":3,"name":"published","label":"Published","group":{"name":"publication_history","label":"Publication History"}}]}}