Pith. sign in

REVIEW 3 major objections 6 minor 75 references

Deep learning enhanced initial model prediction in elastic FWI: application to marine streamer data

T0 review · 3 major / 6 minor · reviewed 2026-08-10 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read Deep learning predicts low-wavenumber starting models for elastic FWI from high-frequency data alone.

desk verdict A solid incremental extension of the authors' earlier CNN-based low-wavenumber prediction to elastic FWI and real data; the central claim survives the benchmarks, but the training/inference mismatch and thin field validation need work. read the letter →

arxiv 2501.12992 v1 pith:PF7DTRSH submitted 2025-01-22 physics.geo-ph

Pavel Plotnitskii (1) , Oleg Ovcharenko (2) , Vladimir Kazei (3) , Daniel Peter (1) , Tariq Alkhalifah (1) ((1) KAUST , (2) NVIDIA , (3) Aramco Americas , Houston
show 1 more author
USA)
This is my paper · ORCID
classification physics.geo-ph
keywords full-waveforminversionelasticFWIdeeplearninglow-wavenumbermodelinitialpredictionmarinestreamerdataFusionNetcycle-skipping
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

Full-waveform inversion (FWI) needs low-frequency seismic data to recover the smooth, low-wavenumber part of the subsurface model, but such data are often missing in marine streamer surveys. This paper tries to show that a convolutional neural network can substitute for that missing information: it maps the high-wavenumber model updates produced by elastic FWI on band-passed data into a low-wavenumber initial model, effectively restoring the background velocity structure. The authors train a FusionNet-based CNN on 3,904 synthetic elastic models built with the field survey's acquisition parameters, then test it on the Marmousi II and Overthrust benchmarks and on a real marine streamer dataset from the North-West Australian shelf. In all three tests, elastic FWI started from the CNN-predicted initial model ends with higher $R^2$ similarity to the true model and fewer cycle-skipping artifacts than FWI started from a one-dimensional initial model. If correct, this offers a practical route to initial velocity model building that does not require low-frequency acquisition, though the paper also documents limitations such as velocity overshoot below 3 km on the field data.

What carries the argument

The load-bearing object is the FusionNet convolutional neural network, a fully residual image-to-image architecture with nested short skip connections inside each residual block in addition to long skip connections. The paper increases the filter count to 110, giving about 231 million trainable parameters, and inserts dropout layers after each convolutional block. The network is fed an 11-channel input: the first ten channels are the successive velocity-model updates from ten elastic FWI iterations run on data high-pass filtered above 5 Hz, and the eleventh channel is a water taper with values from 0 to 1 marking the water column. The target is the low-wavenumber model update, obtained by subtracting the initial model from a Gaussian-smoothed true model. The architecture's cumulative receptive field is what lets shallow layers see high-frequency reflectivity patterns while deeper layers integrate them into smooth background-velocity updates; the water taper supplies the known sea-floor geometry that anchors the low-wavenumber reconstruction.

What would settle it

Take a field dataset from a different basin with unknown water-bottom geometry and salt-containing geology, generate the ten FWI updates on data above 5 Hz, apply the trained network, and compare the resulting FWI convergence and well-log match against a 1D-initial-model baseline; a failure to improve (or a larger final misfit) would falsify the claim that the CNN-predicted initial model generalizes to new marine streamer data.

Watch

Extended reading notes

Core claim

The central claim is that the smooth, low-wavenumber model update that would normally come from low-frequency seismic data can be learned from the high-wavenumber FWI updates obtained on data with low frequencies removed. The authors construct a supervised regression problem in which the input is a stack of the first ten elastic FWI velocity-model updates plus a water taper that encodes the known sea-floor geometry, and the target is the low-wavenumber update obtained by subtracting the initial model from a smoothed true model. A FusionNet-based fully residual convolutional network trained on 3,904 synthetic layered models, generated with the acquisition design of a real marine streamer survey, learns this mapping. On the Marmousi II and Overthrust test models, which were not in the training set, the CNN-predicted initial model improved the final FWI $R^2$ (from 0.569 to 0.694 for Marmousi and from 0.451 to 0.526 for Overthrust) relative to starting from the a priori 1D model, although it remained below the ideal case of starting from a smoothed true model. On the field data, FWI from the CNN-predicted initial model produced fewer cycle-skipping artifacts and better matched the available well log at several depths than FWI from the 1D initial model, with the paper explicitly noting velocity overshoot below 3 km.

Load-bearing premise

The method works only if the synthetic models used for training are representative enough of the real subsurface and acquisition, and if the sea-floor geometry is known; the paper's own field result showing velocity overshoot below 3 km suggests this can fail outside the training distribution.

Editorial extensions

If this is right

  • FWI workflows that currently rely on low-frequency acquisition or heavy regularization can instead initialize from a CNN-predicted low-wavenumber model, reducing the need for frequencies below 5 Hz.
  • A network trained for one marine streamer survey design can be applied to new data from the same survey without retraining, provided the acquisition parameters match.
  • The method complements existing initial-model building tools such as ray-based tomography and gradient filtering, since it operates on FWI by-products and can be inserted into existing workflows.
  • The gain is largest when the a priori 1D initial model is far from the smoothed true model; when the starting 1D model is already accurate, the CNN adds little.
  • Generating the training dataset is computationally expensive (about 20–40 minutes per model on 160 CPU cores), but inference on a new model is fast once the network is trained.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • One implication the authors leave implicit is that the same input-target construction could be carried over to other inverse problems where high-wavenumber updates are available but smooth background structure is missing, such as medical ultrasound tomography or ground-penetrating radar, as long as a training distribution can be sampled.
  • Because the input includes a water taper with known sea-floor geometry, the method's field transfer would be challenged by surveys with unknown or rapidly varying water-bottom topography; an extension would be to train with randomized tapers or predict the water bottom jointly.
  • The velocity overshoot below 3 km on the field data suggests the synthetic generator under-represents deep high-velocity geology; extending the generator with deeper high-velocity layers or salt-like bodies could reduce this bias and improve the field result.
  • The fact that FWI from the smoothed-true initial model still clearly outperforms FWI from the CNN-predicted model (0.870 vs 0.694 on Marmousi) indicates that the remaining bottleneck is the CNN's error relative to the smoothed true model, so further gains would come from better low-wavenumber prediction rather than from changing the FWI scheme.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 6 minor

Summary. This paper proposes a deep-learning workflow for constructing low-wavenumber initial models for elastic full-waveform inversion (FWI) when low-frequency data are missing. A FusionNet-based CNN is trained on 3,904 synthetic random layered models to map ten consecutive high-wavenumber FWI model updates (plus a water taper) into a low-wavenumber model update; the target is a Gaussian-smoothed true model minus the 1D initial model. The trained network is then applied to the Marmousi II and Overthrust benchmark models and to a 2D marine streamer dataset from the Australian shelf. The authors report that FWI starting from the CNN-predicted initial models yields higher R2 scores than the 1D initial model on both benchmarks (Marmousi: 0.653 to 0.694 vs 0.569 to 0.508 for the 1D start; Overthrust: 0.497 to 0.526 vs 0.451 to 0.458), and qualitatively better matches to a single well log on field data, with fewer visible cycle-skipping artifacts.

Significance. Assuming the results hold, the method offers a practical, comparatively cheap way to build initial models for elastic FWI without low frequencies, because the CNN operates on FWI by-products and inference is fast once the network is trained. The paper is transparent about data generation, architecture, and training details, and it validates on two unseen benchmark models; the authors also promise to release code on GitHub. However, the significance is tempered by two weaknesses: the CNN is applied under an FWI configuration (global-correlation misfit, 6/6/7/8/8 Hz stages) that differs from the training configuration (L2 misfit, 8/10 Hz stages) without a sensitivity analysis, and the field-data demonstration relies on a single well log with no quantitative metric. These issues, rather than the core idea, are the main limitations.

major comments (3)
  1. [FWI Results / Wave propagation] The CNN is trained on input updates generated with the L2 misfit and two low-pass stages at 8 Hz and 10 Hz, as described in the 'Wave propagation' section, but at inference the paper uses a global-correlation norm misfit and five stages with low-pass cutoffs 6, 6, 7, 8, and 8 Hz, as described in the first paragraph of 'FWI Results'. The network is thus applied to input updates with different amplitude scaling and spectral content than its training distribution, and no experiment or analysis quantifies its robustness to this shift. Since all benchmark results in Figures 10 and 11 were obtained under the shifted configuration, they show empirical robustness in two cases but do not establish that the learned mapping is the intended one under the exact conditions of use. Please add a sensitivity test, for example by generating CNN inputs with the inference workflow and comparing predictions, or by retraining on the inference configuration.
  2. [Field data] The field-data validation rests entirely on a single acoustic Vp well log and visual inspection; no quantitative metric (such as R2 or correlation along the well) is reported, and no error bars or training-seed variation are provided. The statement in the Conclusions that the method 'led to better FWI results' on the field dataset is therefore not quantitatively supported. Please add a quantitative well-log comparison and, ideally, a small ensemble of training runs to assess variability.
  3. [Overthrust / Figure 11] The R2 improvement from the CNN-predicted initial model on the Overthrust benchmark is 0.497 to 0.526, while the CNN-target initial model (the smoothed true model) actually decreases R2 from 0.678 to 0.674 after FWI. This shows that even the 'ideal' initial model does not guarantee improvement on this benchmark, so the modest gain from the CNN prediction may be within the noise of the workflow. Please discuss this result explicitly and provide additional evidence that the improvement is statistically meaningful, or temper the claim accordingly.
minor comments (6)
  1. [Training dataset preparation / Figure 5] Figure 5's caption says '12 input channels' but the text describes 11 channels (10 FWI updates plus the water taper); please correct the inconsistency.
  2. [Training dataset preparation] The outlier removal step is described only qualitatively ('we removed those models from the training dataset'); please specify the quantitative criterion used, such as the amplitude threshold on the model updates.
  3. [Synthetic data] The Gaussian smoothing kernel used to construct the CNN target is not specified; please report the kernel width or standard deviation.
  4. [Testing CNN performance on Marmousi/Overthrust] In Figures 7 and 8, the subplot labels (b) and (c) are duplicated (one for the CNN inference result and one for the target); please renumber the subplots.
  5. [Conclusions] The conclusion that the method 'led to better FWI results' on field data is stronger than the evidence in the 'Field data' section supports, given the single-well, qualitative comparison; please align the wording with the evidence.
  6. [Conclusions / GitHub statement] The manuscript states that open-access data to reproduce the paper 'will be available' on GitHub; please clarify the current status of the repository and provide a version or commit identifier for reproducibility.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the CNN is trained on synthetic FWI updates with targets derived from true models, then tested on Marmousi, Overthrust, and field data outside the training set.

full rationale

The central claim is an empirical generalization claim: a FusionNet trained on synthetic high-wavenumber FWI updates predicts low-wavenumber initial models that improve FWI. The inputs (10 high-wavenumber Vp updates plus water taper) and targets (smoothed true model minus initial model) are generated independently by the authors' own elastic FWI code; the targets are not defined in terms of CNN outputs. All benchmark tests (Marmousi II, Overthrust) are explicitly 'unrelated'/'not included in the training' models, and the field data is external. The self-citations (Plotnitskii et al. 2019, 2020; Ovcharenko et al. 2021b) are historical and implementation references, not load-bearing mathematical premises; no uniqueness theorem is invoked to rule out alternatives. The training/inference distribution shift (L2 vs global correlation misfit, 8/10 Hz vs 6/6/7/8/8 Hz stages) is a generalization-risk concern, not a circularity, because the CNN is not fitted to the benchmark or field outcomes. No equation in the paper reduces a predicted quantity to a fitted input by construction. Score 0.

Assumptions & free parameters 5 free parameters · 5 assumptions · 0 invented entities

The method introduces no new physical entities; all components are standard seismic inversion and deep learning elements. The primary assumptions are about the relationship between FWI updates and low-wavenumber models, the representativeness of synthetic training data, and the fixed elastic parameter relations.

free parameters (5)
  • FWI stage high-pass filter cutoffs = 6, 6, 7, 8, 8 Hz
    Hand-chosen multiscale schedule for FWI experiments; the same schedule is used for all initial models, so it does not bias the comparison, but the exact values are arbitrary and may not be optimal.
  • FWI gradient damping schedule = 1.5, 1, 0.1, 0.25, 0.125
    Spatial filter damping parameters chosen by hand to stabilize multiscale FWI; constant across experiments.
  • Number of FWI iterations for CNN inputs = 10
    Chosen because results appear sufficient; not justified by a convergence criterion.
  • Gaussian smoothing kernel for target = unspecified
    The target is a Gaussian-smoothed true model, but the kernel width is not stated; the target definition directly shapes what the network learns.
  • Outlier removal threshold = unspecified
    Models producing very small-amplitude high-wavenumber updates were removed from training; the threshold is not quantified, introducing potential selection bias.
assumptions (5)
  • domain assumption High-wavenumber FWI updates contain recoverable low-wavenumber information.
    The method's premise: the CNN can map from high-frequency updates to low-wavenumber updates. Invoked throughout, especially in the Method Overview.
  • domain assumption Synthetic random layered models with elastic transforms are representative of the field geology.
    The training set is built from folded 1D layered models with random gradients; the authors argue this matches the relatively flat Australian shelf geology. Section Datasets.
  • domain assumption Gardner's relation (rho = 0.3 Vp^0.25) and Vp/Vs = sqrt(3) accurately relate elastic parameters.
    Used to generate density and shear velocity from P-wave velocity for both synthetic and field applications.
  • domain assumption The water bottom is known and provided as a water taper input.
    The network requires the water mask; the authors assume the seafloor geometry is known in practice. Section Datasets.
  • ad hoc to paper A Gaussian-smoothed true model minus the initial model is the correct target low-wavenumber update.
    This target choice defines what the network learns; no justification is given that it matches the low-wavenumber update a low-frequency FWI would produce.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Deep learning enhanced initial model prediction in elastic FWI: application to marine streamer data." pith.science (2026). https://pith.science/paper/PF7DTRSH

@misc{pith2026250112992,
  author       = {Pith},
  title        = {Pith review of: Deep learning enhanced initial model prediction in elastic FWI: application to marine streamer data},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/PF7DTRSH}},
  note         = {Machine review of arXiv:2501.12992}
}
read the original abstract

Low-frequency data are essential to constrain the low-wavenumber model components in seismic full-waveform inversion (FWI). However, due to acquisition limitations and ambient noise it is often unavailable. Deep learning (DL) can learn to map from high frequency model updates of elastic FWI to a low-wavenumber model update, producing an initial model estimation as if it was available from low-frequency data. We train a FusionNET-based convolutional neural network (CNN) on a synthetic dataset to produce an initial low-wavenumber model from a set of model updates produced by FWI on the data with missing low frequencies. We validate this DL-fused approach using a synthetic benchmark with data generated in an unrelated model to the training dataset. Finally, applying our trained network to estimate an initial low-wavenumber model based on field data, we see that elastic FWI starting from such a 'DL-fused' model update shows improved convergence on real-world marine streamer data.

Figures

Figures reproduced from arXiv: 2501.12992 by the authors.

Figure 1
Figure 1. Workflow for low-wavenumber extrapolation by [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Field data: a) real shot gather, b) real shot [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Examples of synthetic Vp velocity models from the training dataset. a ”true” velocity model and a 1D velocity trend from it that can be used as initial model for FWI. To construct density, ρ, and shear velocity, Vs, from the generated Vp model values, we use Gardner’s relation (Gardner et al., 1974) ρ = α V β p (1) with α = 0.3, β = 0.25 and a ratio Vp/Vs = √ 3 for Poisson solids. In [PITH_FULL_IMAGE:figures/full_f… view at source ↗
Figures from the paper (9 more)
Figure 4
Figure 4. Figure 4: Well-logs extracted from the field dataset (blue), [PITH_FULL_IMAGE:figures/full_fig_p004_4.png]
Figure 5
Figure 5. Figure 5: Data available for DL training (12 input chan [PITH_FULL_IMAGE:figures/full_fig_p005_5.png]
Figure 6
Figure 6. Figure 6: Histograms of velocity perturbations in the in [PITH_FULL_IMAGE:figures/full_fig_p005_6.png]
Figure 7
Figure 7. Figure 7: CNN inference on the Marmousi testing sample: [PITH_FULL_IMAGE:figures/full_fig_p007_7.png]
Figure 9
Figure 9. Figure 9: CNN inference on field data testing sample: (a) [PITH_FULL_IMAGE:figures/full_fig_p008_9.png]
Figure 10
Figure 10. Figure 10: High-frequency band FWI on Marmousi data. Left column - initial models for FWI, center column - FWI [PITH_FULL_IMAGE:figures/full_fig_p009_10.png]
Figure 11
Figure 11. Figure 11: High-frequency band FWI on Overthrust data. Left column - initial models for FWI, center column - FWI [PITH_FULL_IMAGE:figures/full_fig_p009_11.png]
Figure 12
Figure 12. Figure 12: High-frequency band FWI on marine streamer data from the 1D initial model. a) 1D initial model for FWI, [PITH_FULL_IMAGE:figures/full_fig_p011_12.png]
Figure 13
Figure 13. Figure 13: High-frequency band FWI on marine streamer data from the CNN-predicted initial model. a) CNN-predicted [PITH_FULL_IMAGE:figures/full_fig_p011_13.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

75 extracted references · 71 canonical work pages

  1. [1]

    Smith, P

    Alali, A., R. Smith, P. Nivlet, A. Bakulin, and T. Alkhalifah, 2021, Time-lapse seismic cross-equalization using temporal convolutional networks: First International Meeting for Applied Geoscience & Energy Expanded Abstracts, SEG, 1440--1444

    work page 2021

  2. [2]

    Alkhalifah, T., 2016, Full-model wavenumber inversion: An emphasis on the appropriate wavenumber continuation: Geophysics, 81 , R89--R98

    work page 2016

  3. [3]

    Sava, 2020, Elastic full-waveform inversion with probabilistic petrophysical model constraints: Geophysics, 85 , R101--R111

    Aragao, O., and P. Sava, 2020, Elastic full-waveform inversion with probabilistic petrophysical model constraints: Geophysics, 85 , R101--R111

    work page 2020

  4. [4]

    Adler, S

    Araya-Polo, M., A. Adler, S. Farris, and J. Jennings, 2020, Fast and accurate seismic tomography via deep learning, in Deep Learning: Algorithms and Applications: Springer, 129--156

    work page 2020

  5. [5]

    Jennings, A

    Araya-Polo , M., J. Jennings, A. Adler, and T. Dahlke, 2018, Deep-learning tomography: The Leading Edge, 37 , 58--66

    work page 2018

  6. [6]

    Baeten, G., J. W. de Maag, R.-E. Plessix, R. Klaassen, T. Qureshi, M. Kleemeyer, F. t. Kroode, and Z. Rujie, 2013, The use of low frequencies in a full-waveform inversion and impedance inversion land seismic case study: Geophysical Prospecting, 61 , 701--711

    work page 2013

  7. [7]

    Jarraya, and F

    Birnie, C., H. Jarraya, and F. Hansteen, 2021, An introduction to distributed training of deep neural networks for segmentation tasks with large seismic data sets: Geophysics, 86 , KS151--KS160

    work page 2021

  8. [8]

    Chen, F., and D. Peter, 2018a, Constructing misfit function for full waveform inversion based on sliced wasserstein distance: EAGE Extended Abstracts, European Association of Geoscientists & Engineers, 1--5

    work page

Show all 75 references
  1. [9]

    --------, 2018b, A misfit function based on entropy regularized optimal transport for full-waveform inversion: SEG Technical Program Expanded Abstracts, SEG, 1314--1318

  2. [10]

    --------, 2022, Deconvolutional double-difference misfit measurements and the application for full-waveform inversion

    2022

  3. [11]

    Peter, and M

    Chen, F., D. Peter, and M. Ravasi, 2022, Cycle-skipping mitigation using misfit measurements based on differentiable dynamic time warping: Geophysics, 87 , R325--R335

    2022

  4. [12]

    Chen, F., D. B. Peter, and M. Ravasi, 2021, Misfit functions based on differentiable dynamic time warping for waveform inversion: First International Meeting for Applied Geoscience & Energy Expanded Abstracts

    2021

  5. [13]

    Alkhalifah, 2012, Application of multi-source waveform inversion to marine streamer data using the global correlation norm: Geophysical Prospecting, 60 , 748--758

    Choi, Y., and T. Alkhalifah, 2012, Application of multi-source waveform inversion to marine streamer data using the global correlation norm: Geophysical Prospecting, 60 , 748--758

    2012

  6. [14]

    V., 1989, Approximation by superpositions of a sigmoidal function: Mathematics of Control, Signals and Systems, 2 , 303--314

    Cybenko, G. V., 1989, Approximation by superpositions of a sigmoidal function: Mathematics of Control, Signals and Systems, 2 , 303--314

    1989

  7. [15]

    Guasch, T

    Esser, E., L. Guasch, T. van Leeuwen , A. Y. Aravkin, and F. J. Herrmann, 2016, Total-variation regularization strategies in full-waveform inversion: arxiv.org

    2016

  8. [16]

    Araya-Polo, J

    Farris, S., M. Araya-Polo, J. Jennings, B. Clapp, and B. Biondi, 2018, Tomography: a deep learning vs full-waveform inversion comparison: First EAGE Workshop on High Performance Computing for Upstream in Latin America, European Association of Geoscientists & Engineers, 1--5

    2018

  9. [17]

    C., and C

    Fehmers, G. C., and C. F. W. H\"ocker, 2003, Fast structural interpretation with structure‐oriented filtering: Geophysics, 68 , 1286--1293

    2003

  10. [18]

    Fomel, S., 2007, Shaping regularization in geophysical-estimation problems: Geophysics, 72 , R29--R36

    2007

  11. [19]

    Gardner, G. H. F., L. W. Gardner, and A. R. Gregory, 1974, Formation velocity and density—the diagnostic basics for stratigraphic traps: GEOPHYSICS, 39 , 770--780

    1974

  12. [20]

    Hu, W., Y. Jin, X. Wu, and J. Chen, 2019, A progressive deep transfer learning approach to cycle-skipping mitigation in fwi, in SEG Technical Program Expanded Abstracts 2019: Society of Exploration Geophysicists, 2348--2352

    2019

  13. [21]

    Jin, Y., W. Hu, X. Wu, and J. Chen, 2018, Learn low wavenumber information in fwi via deep inception based convolutional networks, in SEG Technical Program Expanded Abstracts 2018: Society of Exploration Geophysicists, 2091--2095

    2018

  14. [22]

    Alkhalifah, 2017, Efficient full waveform inversion using the excitation representation of the source wavefield : Geophysical Journal International, 210 , 1581--1594

    Kalita, M., and T. Alkhalifah, 2017, Efficient full waveform inversion using the excitation representation of the source wavefield : Geophysical Journal International, 210 , 1581--1594

    2017

  15. [23]

    Kazei, Y

    Kalita, M., V. Kazei, Y. Choi, and T. Alkhalifah, 2018, Regularized full-waveform inversion for salt bodies: SEG Technical Program Expanded Abstracts, SEG, 1043--1047

    2018

  16. [24]

    --------, 2019, Regularized full-waveform inversion with automated salt-flooding: Geophysics, 0 , 1--74

    2019

  17. [25]

    Kalita, and T

    Kazei, V., M. Kalita, and T. Alkhalifah, 2017, Salt-body Inversion with Minimum Gradient Support and Sobolev Space Norm Regularizations : Presented at the 79th EAGE Conference and Exhibition 2017

    2017

  18. [26]

    Ovcharenko, T

    Kazei, V., O. Ovcharenko, T. Alkhalifah, and F. Simons, 2019, Realistically textured random velocity models for deep learning applications: Presented at the 81st EAGE Conference and Exhibition 2019

    2019

  19. [27]

    Ovcharenko, P

    Kazei, V., O. Ovcharenko, P. Plotnitskii, D. Peter, T. Alkhalifah, I. Silvestrov, A. Bakulin, and P. Zwartjes, 2020, Elastic near-surface model estimation from full waveforms by deep learning: SEG Technical Program Expanded Abstracts, SEG, 3872--3876

    2020

  20. [28]

    Ovcharenko, P

    Kazei, V., O. Ovcharenko, P. Plotnitskii, D. Peter, X. Zhang, and T. Alkhalifah, 2021, Mapping full seismic waveforms to vertical velocity profiles by deep learning : Geophysics, R711--R721

    2021

  21. [29]

    Tessmer, and T

    Kazei, V., E. Tessmer, and T. Alkhalifah, 2016, Scattering angle-based filtering via extension in velocity, in SEG Technical Program Expanded Abstracts 2016: Society of Exploration Geophysicists , 1157--1162

    2016

  22. [30]

    Min, and C

    Kim, Y., D.-J. Min, and C. Shin, 2011, Frequency-domain reverse-time migration with source estimation: Geophysics, 76 , S41--S49

    2011

  23. [31]

    P., and J

    Kingma, D. P., and J. Ba, 2014, Adam: A method for stochastic optimization: arXiv preprint arXiv:1412.6980

    2014 arXiv

  24. [32]

    K \"o hn, D., 2011, Time domain 2d elastic full waveform tomography: PhD thesis, Kiel University

    2011

  25. [33]

    De Nil, A

    K\"ohn, D., D. De Nil, A. Kurzmann, A. Przebindowska, and T. Bohlen, 2012, On the influence of model parametrization in elastic full waveform tomography: Geophysical Journal International, 191 , 325--345

    2012

  26. [34]

    Campbell, and J

    Lecomte, J.-C., E. Campbell, and J. Letouzey, 1994, Building the seg/eaeg overthrust velocity macro model: Presented at the EAGE/SEG Summer Workshop, European Association of Geoscientists & Engineers

    1994

  27. [35]

    Amazonas, D

    Lewis, W., D. Amazonas, D. Vigh, and R. Coates, 2014, Geologically constrained full-waveform inversion using an anisotropic diffusion based regularization scheme: application to a 3d offshore brazil dataset: Presented at the SEG Technical Program Expanded Abstracts

    2014

  28. [36]

    Alkhalifah, and Z

    Li, Y., T. Alkhalifah, and Z. Zhang, 2021, Deep-learning assisted regularized elastic full waveform inversion using the velocity distribution information from wells : Geophysical Journal International, 226 , 1322--1335

    2021

  29. [37]

    E., and L

    Li, Y. E., and L. Demanet, 2016, Full-waveform inversion with extrapolated low-frequency data: Geophysics, 81 , R339--R348

    2016

  30. [38]

    Luo, Y., and G. T. Schuster, 1991a, Wave-equation traveltime inversion: Geophysics, 56 , 645--653

  31. [39]

    --------, 1991b, Wave‐equation traveltime inversion: Geophysics, 56 , 645--653

  32. [40]

    Martin, G. S., R. Wiley, and K. J. Marfurt, 2006, Marmousi2: An elastic upgrade for marmousi: The Leading Edge, 25 , 156--166

    2006

  33. [41]

    Brossier, Q

    M \'e tivier, L., R. Brossier, Q. M \'e rigot, E. Oudet, and J. Virieux, 2016, Measuring the misfit between seismograms using an optimal transport distance: application to full waveform inversion : Geophysical Journal International, 205 , 345--377

    2016

  34. [42]

    Baumstein, and E

    Ovcharenko, O., A. Baumstein, and E. Neumann, 2021a, Surface-related multiple elimination through orthogonal encoding in the latent space of convolutional autoencoder: First International Meeting for Applied Geoscience & Energy Expanded Abstracts, EAGE, 1355--1359

  35. [43]

    Kazei, M

    Ovcharenko, O., V. Kazei, M. Kalita, D. Peter, and T. Alkhalifah, 2019a, Deep learning for low-frequency extrapolation from multioffset seismic data: Geophysics, 84 , R989--R1001

  36. [44]

    Kazei, D

    Ovcharenko, O., V. Kazei, D. Peter, and T. Alkhalifah, 2017, Neural network based low-frequency data extrapolation: Presented at the 3rd SEG FWI Workshop: What Are We Getting

    2017

  37. [45]

    --------, 2019b, Style transfer for generation of realistically textured subsurface models: SEG Technical Program Expanded Abstracts, SEG, 2393--2397

  38. [46]

    --------, 2021b, Transferring elastic low frequency extrapolation from synthetic to field data: EAGE expanded abstracts, European Association of Geoscientists & Engineers, 1--5

  39. [47]

    Kazei, D

    Ovcharenko, O., V. Kazei, D. Peter, I. Silvestrov, A. Bakulin, and T. Alkhalifah, 2021c, Dual-band generative learning for low-frequency extrapolation in seismic land data: First International Meeting for Applied Geoscience & Energy Expanded Abstracts, EAGE expanded abstra...

  40. [48]

    Dahl, 2019, Velocity model building from raw shot gathers using machine learning: Presented at the 81st EAGE Conference and Exhibition 2019

    ye, O., and E. Dahl, 2019, Velocity model building from raw shot gathers using machine learning: Presented at the 81st EAGE Conference and Exhibition 2019

    2019

  41. [49]

    Alkhalifah, O

    Plotnitskii, P., T. Alkhalifah, O. Ovcharenko, and V. Kazei, 2019, Seismic model low wavenumber extrapolation by a deep convolutional neural network: ASEG Extended Abstracts, 2019 , 1--5

    2019

  42. [50]

    Kazei, O

    Plotnitskii, P., V. Kazei, O. Ovcharenko, D. Peter, and T. Alkhalifah, 2020, Extrapolation of low wavenumbers in fwi gradients by a deep convolutional neural network: EAGE expanded abstracts, 2020 , 1--5

    2020

  43. [51]

    Quan, T. M., D. G. C. Hildebrand, and W.-K. Jeong, 2021, Fusionnet: A deep fully residual convolutional neural network for image segmentation in connectomics: Frontiers in Computer Science, 3

    2021

  44. [52]

    Ravasi, M., and C. E. Birnie, 2021, A joint inversion-segmentation approach to assisted seismic interpretation

    2021

  45. [53]

    Operto, L

    Ravaut, C., S. Operto, L. Improta, J. Virieux, A. Herrero, and P. Dell'Aversana, 2004, Multiscale imaging of complex structures from multifold wide-aperture seismic data by frequency-domain full-waveform tomography: Application to a thrust belt: Geophysical Journal Internation...

    2004

  46. [54]

    Richardson, A., 2018, Seismic full-waveform inversion using deep learning tools and techniques: arXiv preprint arXiv:1801.07232

    2018 arXiv

  47. [55]

    Fischer, and T

    Ronneberger, O., P. Fischer, and T. Brox, 2015, U-net: Convolutional networks for biomedical image segmentation: International Conference on Medical image computing and computer-assisted intervention, Springer, 234--241

    2015

  48. [56]

    Ahmed, A

    Shen, X., I. Ahmed, A. Brenders, J. Dellinger, J. Etgen, and S. Michell, 2018, Full-waveform inversion: The next leap forward in subsalt imaging: The Leading Edge, 37 , 67b1--67b6

    2018

  49. [57]

    Simard, P. Y., D. Steinkraus, J. C. Platt, et al., 2003, Best practices for convolutional neural networks applied to visual document analysis.: Presented at the Icdar

    2003

  50. [58]

    Ravaut, O

    Skopintseva, L., C. Ravaut, O. Pedersen, F. Maao, and K. Hartvigsen, 2019, Regularization in full waveform inversion: Obc vs streamer data results over a gas cloud: EAGE expanded abstracts, European Association of Geoscientists & Engineers, 1--5

    2019

  51. [59]

    Dowle, 2010, Variable-depth streamer -- a broadband marine solution: First Break, 28

    Soubaras, R., and R. Dowle, 2010, Variable-depth streamer -- a broadband marine solution: First Break, 28

    2010

  52. [60]

    Dowle, and R

    Soubaras, R., R. Dowle, and R. Sablon, 2012, Broadseis : Enhancing interpretation and inversion with broadband marine seismic: Presented at the CSEG Recorder

    2012

  53. [61]

    Alkhalifah, 2019, Robust full-waveform inversion with radon-domain matching filter: Geophysics, 84 , 1--121

    Sun, B., and T. Alkhalifah, 2019, Robust full-waveform inversion with radon-domain matching filter: Geophysics, 84 , 1--121

    2019

  54. [62]

    --------, 2020, Ml-descent: An optimization algorithm for full-waveform inversion using machine learning: Geophysics, 85 , R477--R492

    2020

  55. [63]

    Demanet, 2018, Low frequency extrapolation with deep learning, in SEG Technical Program Expanded Abstracts 2018: Society of Exploration Geophysicists , 2011--2015

    Sun, H., and L. Demanet, 2018, Low frequency extrapolation with deep learning, in SEG Technical Program Expanded Abstracts 2018: Society of Exploration Geophysicists , 2011--2015

    2018

  56. [64]

    --------, 2020, Extrapolated full-waveform inversion with deep learning: Geophysics, 85 , R275--R288

    2020

  57. [65]

    --------, 2021, Deep learning for low-frequency extrapolation of multicomponent data in elastic fwi: IEEE Transactions on Geoscience and Remote Sensing, 60 , 1--11

    2021

  58. [66]

    Brossier, L

    Trinh, P., R. Brossier, L. M\'etivier, J. Virieux, and P. Wellington, 2017, Bessel smoothing filter for spectral-element mesh: Geophysical Journal International, 209 , 1489--1512

    2017

  59. [67]

    van Leeuwen , T., and F. J. Herrmann, 2013, Mitigating local minima in full-waveform inversion by expanding the search space: Geophysical Journal International, 195 , 661--667

    2013

  60. [68]

    Operto, 2009, An overview of full-waveform inversion in exploration geophysics: Geophysics, 74 , WCC1--WCC26

    Virieux, J., and S. Operto, 2009, An overview of full-waveform inversion in exploration geophysics: Geophysics, 74 , WCC1--WCC26

    2009

  61. [69]

    Ratcliffe, T

    Warner, M., A. Ratcliffe, T. Nangoo, J. Morgan, A. Umpleby, N. Shah, V. Vinje, I. Stekl, L. Guasch, C. Win, G. Conroy, and A. Bertrand, 2013, Anisotropic 3D full-waveform inversion: Geophysics, 78 , R59--R80

    2013

  62. [70]

    Wu, X., Z. Geng, Y. Shi, N. Pham, S. Fomel, and G. Caumon, 2020, Building realistic structure models to train convolutional neural networks for seismic structural interpretation: Geophysics, 85 , WA27--WA39

    2020

  63. [71]

    Wu, Y., and G. A. McMechan, 2020, Cnn-boosted full-waveform inversion: SEG Technical Program Expanded Abstracts 2020, SEG, 1526--1530

    2020

  64. [72]

    Ma, 2019, Deep-learning inversion: a next generation seismic velocity-model building method: Geophysics, 84 , 1--133

    Yang, F., and J. Ma, 2019, Deep-learning inversion: a next generation seismic velocity-model building method: Geophysics, 84 , 1--133

    2019

  65. [73]

    da Silva, and D

    Yao, G., N. da Silva, and D. Wu, 2019, Reflection-waveform inversion regularized with structure-oriented smoothing shaping: Pure Appl. Geophys., 176 , 5315--5335

    2019

  66. [74]

    Alkhalifah, 2019, Regularized elastic full-waveform inversion using deep learning: Geophysics, 84 , R741--R751

    Zhang, Z.-D., and T. Alkhalifah, 2019, Regularized elastic full-waveform inversion using deep learning: Geophysics, 84 , R741--R751

    2019

  67. [75]

    write newline

    " write newline "" before.all 'output.state := FUNCTION fin.entry add.period write newline FUNCTION new.colon output.state before.all = 'skip set.colon 'output.state := if FUNCTION new.block output.state before.all = 'skip after.block 'output.state := if FUNCTION new.sentence ...

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.