README.md

---
license: cc
task_categories:
- object-detection
size_categories:
- n<1K
---

# Text Region Detection in Historical Astronomical Diagrams 

Official repository of the paper _"Text region detection in historical astronomical diagrams"_. We introduce the first large, diverse, open-access dataset of **948** historical astronomical diagrams annotated with **10,940** oriented polygonal text regions that spans ten centuries (8<sup>th</sup> to 18<sup>th</sup>) and seven major traditions: Arabic, Persian, Chinese, Byzantine, Latin, Hebrew, and Sanskrit.

# Dataset 📜
## Content
We provide our dataset under two directories, namely `EIDA` and `EIDALatin`, with annotations in `LabelMe` format. `EIDA` contains images and associated annotations of all traditions (including Latin) under `train`, `val`, and `test` splits:
```
EIDA/
├── train/
│   ├── <filename_1>.jpg
│   ├── <filename_1>.json
│   │   .
│   │   .
│   ├── <filename_N>.jpg
│   └── <filename_N>.json
├── val/
└── test/
```

For class-aware annotations, we provide `EIDALatin`, which contains Latin subset with text classes, and splits in `.txt` format:
```
EIDALatin/
├── data/
│   ├── <filename_1>.jpg
│   ├── <filename_1>.json
│   │   .
│   │   .
│   ├── <filename_N>.jpg
│   └── <filename_N>.json
├── train.txt
├── val.txt
└── test.txt
```

# Citation 🔖
```
@inproceedings{baltaci2026text,
  title={Text region detection in historical astronomical diagrams},
  author={Baltaci, Zeynep Sonat and Baena, Rapha\"el and Meng, Fei and Norindr, Som and Somer, Florence and Husson, Matthieu and Aubry, Mathieu},
  booktitle={ICDAR},
  year={2026}
}
```

# License
[![License: CC BY 4.0](https://img.shields.io/badge/License-CC%20BY%204.0-lightgrey.svg)](https://creativecommons.org/licenses/by/4.0/)

This work is licensed under a [Creative Commons Attribution 4.0 International License](https://creativecommons.org/licenses/by/4.0/).

# Acknowledgements
This work was funded by the ANR project EIDA ANR-22-CE38-0014, the ANR project VHS ANR-21-CE38-0008, and the ERC project DISCOVER funded by the European Union’s Horizon Europe Research and Innovation program under grant agreement No. 101076028. This work was granted access to the HPC resources of IDRIS under the allocation AD010614956R1 and AD011015222 made by GENCI. The authors would like to thank the many historians and computer vision researchers that contributed to the development of the dataset: Eleonora Andriani (Sphaera project, Max Planck Institute for the History of Sciences, Berlin), Ji Chen, Samuel Guessner, Divna Manolova, Scott Trigg (EIDA project), Malamatenia Vlachou Efstathiou, Léore Bensabath (ENPC), and Vidal Attias (CEA).