U4RASD/omar-al-saleh-manuscripts-segments

Omar Al-Saleh Manuscripts — Segments

Line-level segmented images with transcriptions from the Omar Al-Saleh memoir collection (1951–1965), part of the NAKBA NLP 2026: Arabic Manuscript Understanding Shared Task.

Dataset

Split	Images	With text
train	15,969	15,969
test	2,095	2,095
blind_test	2,671	2,671

Each example contains:

• image: A cropped line image from a manuscript page (JPG or PNG)
• text: The Arabic transcription of that line
• filename: Original filename

Source

The manuscripts are handwritten Arabic memoirs by Omar Al-Saleh spanning 16 documents (~1.6M words, ~6,395 pages) covering the period 1951–1965. Line-level segmentation and expert-verified transcriptions were produced for this shared task.

Usage

from datasets import load_dataset

ds = load_dataset("U4RASD/omar-al-saleh-manuscripts-segments")
example = ds["train"][0]
example["image"].show()
print(example["text"])

Citation

If you use this dataset, please cite the NAKBA NLP 2026 shared task:

@inproceedings{nakbanlp-ar-ms,
  title     = {NAKBA NLP 2026: Shared Task on Arabic Handwritten Manuscript Understanding (Palestine Memory–Omar Al-Saleh Memoir)},
  author    = {Hamoud, Hadi and Chamseddine, Ahmad Ali and Shalash, Bilal and
               Abid, Firas Ben and Jarrar, Mustafa and Chakra, Chadi Abou and
               Ghanem, Bernard and Zaraket, Fadi A.},
  booktitle = {Proceedings of the 2nd International Workshop on Nakba Narratives as Language Resources (Nakba-NLP 2026), co-located with the Language Resources and Evaluation Conference (LREC 2026)},
  year      = {2026},
  address   = {Palma, Mallorca, Spain},
  month     = {May}
}

Contact

ar-ms@dohainstitute.edu.qa, hhamoud@dohainstitute.edu.qa