dataset_info:
features:
- name: image
dtype: image
- name: alto
dtype: string
- name: filename
dtype: string
splits:
- name: train
num_bytes: 1237612382
num_examples: 755
- name: val
num_bytes: 1191110473
num_examples: 755
- name: test
num_bytes: 2378929906
num_examples: 1511
download_size: 4702203923
dataset_size: 4807652761
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: val
path: data/val-*
- split: test
path: data/test-*
tags:
- document-layout-analysis
- historical-documents
- icdar-competition
pretty_name: ICDAR 2019 Competition on Baseline Detection (cBAD)
license: cc-by-4.0
ICDAR 2019 Competition on Baseline Detection (cBAD) Dataset
Dataset Description
The ICDAR 2019 Competition on Baseline Detection (cBAD) dataset was developed for the eponymous competition focused on Baseline Detection in historical documents. Baseline detection is a critical step for segmenting handwritten lines before Handwritten Text Recognition (HTR).
This is a newly created, real-world dataset consisting of 3,021 annotated document page images collected from seven different European archives.
Dataset Creation
Source Data
The document images were collected from various archival sources across Europe as part of the READ project (Recognition and Enrichment of Archival Documents).
Creators
The primary creators of the competition and dataset are: Markus Diem, Florian Kleber, and Basilis Gatos.
Funding
The dataset was created under the auspices of the European Commission's READ (Recognition and Enrichment of Archival Documents) project (Grant ID: 674943).
License
The dataset is licensed under the Creative Commons Attribution 4.0 International (CC-BY-4.0) license.
Citation
When using this dataset, please cite the original Zenodo record as follows:
@dataset{DiemKleberGatos2019,
author = {Diem, Markus and Kleber, Florian and Gatos, Basilis},
title = {{ICDAR 2019 Competition on Baseline Detection (cBAD)}},
publisher = {Zenodo},
year = {2019},
doi = {10.5281/zenodo.3568023},
url = {[https://zenodo.org/records/3568023](https://zenodo.org/records/3568023)}
}