Enhanced Parquet Edition: COVERAGE: A Novel Database for Copy-Move Forgery Detection

This repository packages the COVERAGE evaluation assets supplied by ForgeryVCR into a Hugging Face Parquet configuration. It contains 200 images: 100 authentic images and 100 copy-move forgeries, with 100 supplied pixel masks. The images and masks are the 832 × 832 preprocessed ForgeryVCR release, embedded as their existing PNG bytes.

Subsets

Config	Split	Rows	Image size	Pixel masks	Parquet files
coverage	full	200	832 × 832	Forged images only	1

The full split contains all supplied authentic and forged COVERAGE samples.

Columns

• image: the supplied RGB image, decoded with the Hugging Face Image feature.
• mask: the supplied grayscale annotation; black 0 marks background and white 255 marks the forged region. Intermediate grayscale values are preserved. Authentic images have mask = null.
• label: image-level authenticity label, 0 = real, 1 = fake.
• manipulation: real or copy-move.
• sample_id, source_image_path: the image path relative to the extracted release root.
• source_mask_path: the supplied mask path relative to the extracted release root; null for authentic images.
• source_split: the upstream ForgeryVCR split, test.
• filename_source_id: the numeric image identifier shared by an authentic image and its forged counterpart.
• width, height: image dimensions in this release, in pixels.
• has_mask: true for the 100 forged images and false for the 100 authentic images.

Metadata Files

Each configuration includes stats.json with split counts, manipulation counts, image dimensions, supplied-mask statistics, and Parquet shard paths and sizes.

The root inventory.json collects these statistics and records the total row count, shard count, and encoded byte sizes. provenance/ retains the supplied image lists, image-to-mask annotations, and upstream README.

Citation

If you reference COVERAGE, please cite the original paper:

@inproceedings{wen2016coverage,
  author={Wen, Bihan and Zhu, Ye and Subramanian, Ramanathan and Ng, Tian-Tsong and Shen, Xuanjing and Winkler, Stefan},
  title={COVERAGE -- A Novel Database for Copy-Move Forgery Detection},
  booktitle={2016 IEEE International Conference on Image Processing (ICIP)},
  pages={161--165},
  year={2016},
  doi={10.1109/ICIP.2016.7532339}
}

Dataset Export

This Parquet edition reorganizes the ForgeryVCR COVERAGE release of the original COVERAGE dataset and adds per-sample metadata. The original data terms restrict use to non-commercial research. Please also reference this dataset repository when using the packaged version.

@misc{thanejoss_coverage_2026,
  author={{ThaneJoss}},
  title={{COVERAGE}},
  year={2026},
  publisher={Hugging Face},
  url={https://huggingface.co/datasets/ThaneJoss/COVERAGE},
  note={Dataset available on Hugging Face; ForgeryVCR preprocessed release}
}