arXiv subjects

Sergey Fomel

Publications and source records attributed to Sergey Fomel.

6 recordsLinked to original sources

arXiv2026

Artificial Intelligence for Subsurface Imaging Understanding: A Decade Review of Challenges, Methods, Benchmarks, and Outlook

Subsurface imaging interpretation bridges observed geophysical data and quantitative geological models, supporting hydrocarbon exploration, CO2 storage assessment, and geohazard monitoring. Over the past decade, machine learning and deep learning have substantially reshaped interpretation workflows. This review synthesizes the 2015-2025 literature across four tasks: structural interpretation, geobody identification, seismic facies analysis, and property estimation, tracing the field's evolution from classical machine learning through deep learning to emerging domain foundation models, and how these tasks couple within a single interpretation system. The task remains fundamentally different from other AI applications, facing ambiguous signals, interpretive non-uniqueness, sparse semantics, unfixed target locations, and scarce reliable annotations. We synthesize three defining challenges: interpretation under complex geological conditions, cross-survey semantic generalization under low information density, and the absence of reliable benchmarks. Addressing them will hinge on integrating human expertise, physical constraints, and geological priors into training and inference, and on treating uncertainty quantification as an intrinsic model output. We outline a forward-looking agenda: unified, jointly modelled interpretation systems with cross-task consistency; priors evolving from physics toward language and multimodal supervision; end-to-end uncertainty propagation; human-AI collaboration and agent-orchestrated workflows; and a more rigorous evaluation science supported by an AI-ready data ecosystem. The review is accompanied by an open benchmark resource (CIG-Bench), covering fault segmentation, relative geologic time estimation, geobody segmentation, and property modeling, with synthetic datasets, pretrained baselines, and quantitative evaluation: https://douyimin.github.io/CIG-bench

physics.geo-ph↗

arXiv2026

Training a generalizable diffusion model for seismic data processing using a large-scale open-source waveform dataset

We introduce the Seismic Waveforms dataset for Automatic Neural-network processing (SWAN), a comprehensive and standardized benchmark designed to advance data-driven seismic signal processing. SWAN aggregates diverse synthetic and real seismic waveforms spanning a wide range of geological structures, noise conditions, propagation environments, and acquisition geometries, providing a unified foundation for training highly generalizable models. Leveraging this dataset, we develop and evaluate a conditionally constrained residual diffusion model for core seismic processing tasks, focusing on missing-trace reconstruction. Extensive experiments demonstrate that diffusion models trained on SWAN achieve state-of-the-art performance across heterogeneous testing scenarios, outperforming leading deep-learning and physics-based baselines on both synthetic benchmarks and field data examples. The results highlight SWAN's value as both a scalable training corpus and a rigorous evaluation framework, and illustrate the strong potential of diffusion-based architectures for robust, generalizable seismic data processing.

physics.geo-ph↗

arXiv2026

Phase variance as a seismic quality-control attribute

Seismic wavefields recorded on land are strongly distorted by near-surface heterogeneity, introducing trace-specific, frequency-dependent phase perturbations that persist even after advanced time processing. Conventional surface-consistent deconvolution targets long- to mid-wavelength phase variability through overdetermination, but cannot correct localized, non-surface-consistent distortions, and its effectiveness degrades when such effects dominate, as is often the case for point-receiver data. Additionally, conventional workflows provide no direct, quantitative measure of phase reliability; phase quality is assessed only indirectly through amplitude behavior or visual inspection, leaving residual phase disorder largely undiagnosed. We introduce phase variance as a seismic quality-control attribute by treating seismic phases as circular random variables and analyzing local trace ensembles using circular statistics. This data-driven measure quantifies localized phase dispersion without phase unwrapping, enabling analysis of local phase trends without global assumptions or wavelet models. Phase variance is computed automatically and provides frequency-by-frequency classification from coherent signal to fully randomized, noise-dominated phase. Synthetic tests confirm that phase variance reliably captures imposed phase perturbations and their frequency dependence. Application to field prestack land data shows that conventional processing reduces phase variability primarily in the low-to-intermediate frequency range and struggles within the noise cone, while the highest and lowest frequencies show little improvement. Phase variance operates automatically over the full prestack volume, frequency by frequency, providing a consistent, human-independent metric for defining effective bandwidth and supporting phase-sensitive workflows such as AVO, migration, and full-waveform inversion.

physics.geo-ph↗

arXiv2025

On the workflow, opportunities and challenges of developing foundation model in geophysics

Foundation models, as a mainstream technology in artificial intelligence, have demonstrated immense potential across various domains in recent years, particularly in handling complex tasks and multimodal data. In the field of geophysics, although the application of foundation models is gradually expanding, there is currently a lack of comprehensive reviews discussing the full workflow of integrating foundation models with geophysical data. To address this gap, this paper presents a complete framework that systematically explores the entire process of developing foundation models in conjunction with geophysical data. From data collection and preprocessing to model architecture selection, pre-training strategies, and model deployment, we provide a detailed analysis of the key techniques and methodologies at each stage. In particular, considering the diversity, complexity, and physical consistency constraints of geophysical data, we discuss targeted solutions to address these challenges. Furthermore, we discuss how to leverage the transfer learning capabilities of foundation models to reduce reliance on labeled data, enhance computational efficiency, and incorporate physical constraints into model training, thereby improving physical consistency and interpretability. Through a comprehensive summary and analysis of the current technological landscape, this paper not only fills the gap in the geophysics domain regarding a full-process review of foundation models but also offers valuable practical guidance for their application in geophysical data analysis, driving innovation and advancement in the field.

physics.geo-ph↗

arXiv2013

Seismic modeling using the frozen Gaussian approximation

We adopt the frozen Gaussian approximation (FGA) for modeling seismic waves. The method belongs to the category of ray-based beam methods. It decomposes seismic wavefield into a set of Gaussian functions and propagates these Gaussian functions along appropriate ray paths. As opposed to the classic Gaussian-beam method, FGA keeps the Gaussians frozen (at a fixed width) during the propagation process and adjusts their amplitudes to produce an accurate approximation after summation. We perform the initial decomposition of seismic data using a fast version of the Fourier-Bros-Iagolnitzer (FBI) transform and propagate the frozen Gaussian beams numerically using ray tracing. A test using a smoothed Marmousi model confirms the validity of FGA for accurate modeling of seismic wavefields.

physics.geo-ph↗

arXiv2008

Fast Computation of Partial Fourier Transforms

We introduce two efficient algorithms for computing the partial Fourier transforms in one and two dimensions. Our study is motivated by the wave extrapolation procedure in reflection seismology. In both algorithms, the main idea is to decompose the summation domain of into simpler components in a multiscale way. Existing fast algorithms are then applied to each component to obtain optimal complexity. The algorithm in 1D is exact and takes $O(N\log^2 N)$ steps. Our solution in 2D is an approximate but accurate algorithm that takes $O(N^2 \log^2 N)$ steps. In both cases, the complexities are almost linear in terms of the degree of freedom. We provide numerical results on several test examples.

math.NA↗