OCR-D Quick Start Guide

Open your Ubuntu Terminal

On Ubuntu, open your Terminal.
On Windows, install WSL, Ubuntu and Docker Desktop by following these steps:

  1. Install WSL 2 by opening the PowerShell and running: 
    wsl --install

  2. Download and install Ubuntu 22.04.2 LTS from Microsoft App Store.

  3. Open Ubuntu 22.04.2 LTS and follow the instructions.

  4. Install Docker Desktop and set it up for WSL 2.

  5. Make sure, Docker Desktop is running.

Install and set up Docker

In the Ubuntu shell, run:

docker ps

If the command is not found, you may need to install Docker first.

Further Requisites

  1. Install OCR-D via Docker and download example data from Github: 
    docker pull ocrd/all:maximum
mk dir ocr-d
cd ocr-d
git clone https://github.com/OCR-D/assets
mkdir workspace
cp -r assets/data/kant_aufklaerung_1784 workspace/kant_aufklaerung_1784
cd workspace/kant_aufklaerung_1784
  2. Set up Docker : 
    sudo systemctl enable docker
sudo usermod -aG docker $USER
docker run --workdir /data --volume $PWD/.config:/.config --volume $PWD:/data --volume $PWD/models:/usr/local/share/ocrd-resources --volume $PWD/models:/usr/local/share/tessdata --volume $PWD/models:/usr/local/share/ocrd-resources -it ocrd/all bash
  3. Download some models: 
    mkdir -p $PWD/models/ocrd-tesserocr-recognize
sudo mkdir -p /usr/local/share/ocrd-resources/ocrd-tesserocr-recognize/configs
cd data
ocrd resmgr download '*'

First minimal workflow with OCR-D

ocrd-tesserocr-recognize -I OCR-D-IMG -O OCR-D-TESSOCR -P segmentation_level region -P textequiv_level word -P find_tables true -P model ocrd-tesserocr-recognize -I OCR-D-IMG -O OCR-D-TESSOCR -P segmentation_level region -P textequiv_level word -P find_tables true -P model Fraktur_GT4HistOCR

Congratulations! You ran your first (minimal) OCR-D Workflow.
You will find the results in the directory workspace/kant_aufklaerung_1784 under data.
Consult the Setup Guide for more details and other installation methods or jump into the User Guide to learn more about OCR‑D. Below you find a short explanation for the ocrd-tesserocr-recognize command.

Explanation

The commandocrd-tesserocr-recognize -I OCR-D-IMG -O OCR-D-TESSOCR -P segmentation_level region -P textequiv_level word -P find_tables true -P model Fraktur_GT4HistOCR for the recognition contains the following parameters:

  1. ocrd-tesserocr-recognize is the processor used.
  2. -I is followed by the name of the input folder, here images.
  3. -O is followed by the name of the output folder where you will find the results (here binarised images and mets files with the recognised text.
  4. -P segmentation_level region is a parameter for the processor which tells tesserocr to start the segmentation on the level of regions.
  5. -P textequiv_level word is a parameter for the processor which tells tesserocr to stop the segmentation on the level of words (meaning glyphs will not be segmented here).
  6. -P find_tables true is a parameter for the processor which tells tesserocr to recognise tables.
  7. -P model Fraktur_GT4HistOCR is a parameter for the processor which tells tesserocr to use the model Fraktur_GT4HistOCR for recognition.