evndttr's picture
Update README.md
4293a9f verified
|
Raw
History Blame Contribute Delete
2.37 kB
metadata
language: fr
license: cc-by-nc-4.0
dataset_info:
  - config_name: default
    features:
      - name: id
        dtype: string
      - name: image
        dtype: image
      - name: label
        dtype: string
      - name: page
        dtype: string
    splits:
      - name: train
        num_examples: 940
        num_bytes: 255688583
size_categories: 1K<n<10K
task_categories:
  - image-to-text
  - other
pretty_name: Manuscrit Moreau 841  Jacques Cartier, *Relation* (1534)
annotations_creators:
  - human-annotated
language_creators:
  - found
tags:
  - early-modern-french
  - manuscript
  - exploration
  - cartier-1534
  - transcription
  - humanities

Manuscrit Moreau 841 — Jacques Cartier, Relation (1534)


Description

This dataset contains a transcription of the manuscript known as Ms. Moreau 841 (Bibliothèque nationale de France), a manuscript copy of the Relation describing Jacques Cartier’s 1534 voyage to North America.

The present dataset was transcribed and aligned to corresponding image segments to support research in handwritten text recognition.

BnF Catalog record: [https://gallica.bnf.fr/ark:/12148/btv1b10035580h/f57/f57.item)


Dataset contents

Field Type Description Example
id string Unique line identifier (page + line number) P001_L001
image string Relative path or URL to cropped line image images/P001_L001.jpg
page string Source page identifier P001
label string Diplomatic transcription of the handwritten line Le vingt & cinq iour d’aoust nous descouvrismes vne terre...

Transcription parameters

  • Original spelling, capitalization, and punctuation preserved.
  • Abbreviations not expanded unless essential for clarity.
  • Lineation follows the manuscript: one record per handwritten line.
  • Ligatures and special characters normalized to Unicode NFKC.

The transcription was completed manually using eScriptorium for line segmentation and annotation, exported in PAGE XML (2019-07-15) format, and then converted into tabular form for reuse and model training.


Use cases

This dataset can support:

  • Training or evaluating OCR/HTR models on early modern French handwriting.
  • Digital humanities and paleography research.