Reproducible material for FreqSiameseFWI: A Novel Deep Learning Framework for Multi-Source Full Wave Inversion - Omar M. Saad and Tariq Alkhalifah
Click here to access the Project Report. Authentication to the Restricted Area filespace is required.
This repository is organized as follows:
- 📂 asset: folder containing logo;
- 📂 data: folder containing Marmousi2 and overthrust models data;
- 📂 utils: set of common function to run FWI;
- 📂 Model: containing FreqSiamese network;
- 📂 results: containing the reconstructed velocity model using FreqSiameseFWI;
The following notebooks are provided:
- 📙
FreqSiameseFWI_Marmousi_FWI.ipynb
: the main notebook performing the FreqSiameseFWI for Marmousi model; - 📙
FreqSiameseFWI_overethrust_FWI.ipynb
: the main notebook performing the FreqSiameseFWI for overthrust model; - 📙
FreqSiameseFWI_Marmousi_MSFWI.ipynb
: the main notebook performing the FreqSiameseFWI for Marmousi model (multi-sources); - 📙
FreqSiameseFWI_overethrust_MSFWI.ipynb
: the main notebook performing the FreqSiameseFWI for overthrust model (multi-sources);
- To ensure the reproducibility of the results, we suggest using the
FWIGAN.yml
file when creating an environment. - Please install the Deepwave 0.0.8 toolbox version, which is used in this project. Run:
./install_env.sh
It will take some time, but if you see the word Done!
on your terminal you are ready to go.
Remember to always activate the environment by typing:
conda activate FWIGAN
Disclaimer: All experiments have been carried on a Intel(R) Xeon(R) CPU @ 2.10GHz equipped with a single NVIDIA GEForce RTX 3090 GPU. Different environment configurations may be required for different combinations of workstation and GPU.
@article{saad2025f,
title={F-SiameseFWI: A Novel Deep Learning Framework for Multi-Source Full Wave Inversion},
author={Saad, Omar M and Alkhalifah, Tariq},
journal={Geophysics},
volume={90},
number={4},
pages={1--51},
year={2025},
doi ={https://doi.org/10.1190/geo2024-0785.1},
publisher={Society of Exploration Geophysicists}
}