Datasets:

LIACC
/

PORTO

dataset_info:
  features:
    - name: corpus
      dtype: string
    - name: book
      dtype: string
    - name: date
      dtype: int32
    - name: filename
      dtype: string
    - name: transcription
      dtype: string
    - name: aadi
      dtype: string
    - name: paddleocr
      dtype: string
    - name: tesseract
      dtype: string
    - name: transkribus
      dtype: string
    - name: img
      dtype: image
    - name: complex_layout
      dtype: bool
    - name: language
      dtype: string
  splits:
    - name: train
      num_examples: 3782
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/PORTO-part-*
license: cc-by-4.0
task_categories:
  - image-to-text
  - fill-mask
  - text-generation
language:
  - pt
tags:
  - OCR
  - Post-OCR
  - Historical
size_categories:
  - 1K<n<10k
pretty_name: >-
  Corpus / Book / Date / Filename / Transcription / AADI / PaddleOCR / Tesseract
  / Transkribus / Image / Complex Layout / Language

Post-OCR Resources for Text Optimisation

Resource for evaluation and develop OCRs and Post-OCR focused on historical Portuguese.

How to load the dataset:

from datasets import load_dataset

dataset = load_dataset("LIACC/PORTO")

Citation

When using or citing this corpus, kindly cite the following publication:

@inproceedings{10.1145/3746252.3761633,
author = {Freitas Os\'{o}rio, Tom\'{a}s and Lopes Cardoso, Henrique},
title = {Portuguese post-OCR Resources for Text Optimisation},
year = {2025},
isbn = {9798400720406},
publisher = {Association for Computing Machinery},
address = {New York, NY, USA},
url = {https://doi.org/10.1145/3746252.3761633},
doi = {10.1145/3746252.3761633},
booktitle = {Proceedings of the 34th ACM International Conference on Information and Knowledge Management},
pages = {6361–6366},
numpages = {6},
keywords = {historical portuguese corpora, ocr evaluation, post-ocr correction resources},
location = {Seoul, Republic of Korea},
series = {CIKM '25}
}