wjbmattingly's picture
Update README.md
6778834 verified
metadata
license: cc-by-4.0
dataset_info:
  features:
    - name: image_id
      dtype: string
    - name: width
      dtype: int32
    - name: height
      dtype: int32
    - name: objects
      struct:
        - name: area
          sequence: int64
        - name: bbox
          sequence:
            sequence: int64
        - name: category
          sequence: int64
        - name: id
          sequence: string
    - name: image
      dtype: image
  splits:
    - name: train
      num_bytes: 351974104
      num_examples: 121
  download_size: 351271785
  dataset_size: 351974104
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*

This is a dataset for doing line detection on medieval manuscripts. The data was collected from the various projects on HTR United.

Sources for the Data

CREMMA Medii Aevi

GitHub HTR-United

@misc{YourReferenceHere,
author = {Clérice, Thibault and Chagué, Alix and Vlachou-Efstathiou, Malamatenia},
doi = {10.5281/zenodo.7013436},
title = {CREMMA Medii Aevi},
url = {https://github.com/HTR-United/CREMMA-Medieval-LAT}
}

Clérice T., Chagué A., Vlachou-Efstathiou M. CREMMA Medii Aevi DOI: 10.5281/zenodo.7013436 URL: https://github.com/HTR-United/CREMMA-Medieval-LAT