Project structure

This repository is organized as follows:

• 📂 asset: folder containing logo;
• 📂 velocity: folder containing Marmousi model;
• 📂 utils: set of common function to run FWI;
• 📂 results/fwi_image: containing the inverted results;
• 📂 Model: containing max-pooling operator;
• 📂 observed_data: containing the synthetic dataset

Notebooks

The following notebooks are provided:

• 📙 Forward.ipynb: notebook performing forward modeling to generate the synthetic dataset;
• 📙 FWI.ipynb: notebook performing MPBAEP-FWI;
• 📙 plot.ipynb: notebook plotting the inverted velocity model

Getting started 👾 🤖

To ensure reproducibility of the results, we suggest using the environment.yml file when creating an environment.

Simply run:

./install_env.sh

It will take some time, if at the end you see the word Done! on your terminal you are ready to go.

Remember to always activate the environment by typing:

conda activate MPBAEPFWI

Disclaimer: All experiments have been carried on a Intel(R) Xeon(R) Silver 4316 CPU @ 2.30GHz equipped with a single NVIDIA A100 GPU. Different environment configurations may be required for different combinations of workstation and GPU.

Cite us

Xinru Mu, Omar M. Saad, Shaowen Wang, and Tariq Alkhalifah. (2026) Mitigating cycle skipping in full waveform inversion using max-pooling-based approximate envelope and shot patching.