Surojit-Utah/adasemseg-seismic-facies-datasets

Processed Seismic Facies Datasets (F3, Parihaka, Penobscot)

Processed, ready-to-train .npy volumes and train/val/test splits for the three public seismic facies datasets used in:

Saha, S. and Whitaker, R. AdaSemSeg: An Adaptive Few-shot Semantic Segmentation of Seismic Facies. IEEE Transactions on Geoscience and Remote Sensing, 2025. arXiv:2501.16760

Code: github.com/Surojit-Utah/AdaSemSeg · Also archived on Zenodo: 10.5281/zenodo.21764042

Important: this is not originally curated data

These are public, third-party seismic surveys — not data collected or originally curated as part of this project. This repository packages them into a common, ready-to-train format (.npy arrays + train/val/test split JSON) for reproducibility of the AdaSemSeg experiments; it is a processed redistribution for empirical evaluation and benchmarking, not a new dataset release. Please cite the original sources below, not just this repository, when using this data:

Penobscot note: the raw source file (dataset.h5) carries no embedded source/license metadata beyond the citation above, so only the processed seismic.npy/seismic_labels.npy derived from it are included here — matching the same choice already made in the main GitHub repo. The exact processing script (reorientation, cropping, and the transformation applied) is at `data/Penobscot/process_data.py` in the main repo, included there for full transparency and reproducibility.

Contents

F3/
├── train/train_seismic.npy, train_labels.npy
├── test/test1_seismic.npy, test1_labels.npy, test2_seismic.npy, test2_labels.npy
└── split_train_val_test_f3.json
Parihaka/
├── parihaka_facies_train_images.npy, parihaka_facies_train_labels.npy
└── split_train_val_test_parihaka.json
Penobscot/
├── seismic.npy, seismic_labels.npy
└── split_train_val_test_penobscot.json

Usage

These datasets are consumed directly by the training/evaluation code in the main GitHub repo. See `scripts/download_assets.py` (which pulls the same data from the Zenodo record) and the dataset layout section of the README for how the code expects these files to be laid out.

Citation

If you use these processed files, please cite both the AdaSemSeg paper and the original dataset source(s) from the table above:

@article{saha2025adasemseg,
  title={AdaSemSeg: An Adaptive Few-shot Semantic Segmentation of Seismic Facies},
  author={Saha, Surojit and Whitaker, Ross},
  journal={IEEE Transactions on Geoscience and Remote Sensing},
  year={2025},
  doi={10.1109/TGRS.2025.3595010}
}

@dataset{saha2026adasemsegdata,
  title={AdaSemSeg: Processed Seismic Facies Datasets (F3, Parihaka, Penobscot)},
  author={Saha, Surojit and Whitaker, Ross},
  year={2026},
  publisher={Zenodo},
  doi={10.5281/zenodo.21764042},
  url={https://doi.org/10.5281/zenodo.21764042}
}