LAREX is a semi-automatic open-source tool for layout analysis on early printed books. It uses a rule based connected components approach which is very fast, easily comprehensible for the user and allows an intuitive manual correction if necessary. The PageXML format is used to support integration into existing OCR workflows. Evaluations showed that LAREX provides an efficient and flexible way to segment pages of early printed books.

Please feel free to visit the tool homepage and the web application. A short user manual is available here.

• Installation
  • Linux
  • Windows
  • Mac OS X
• Usage
• Configuration
• Related Publications

Additional information about developing for LAREX see here

This guide uses tomcat 8 and Ubuntu (please adjust accordingly for your setup)

• Install required packages: apt-get install tomcat8 maven openjdk-8-jdk
• Clone Repository: git clone https://github.com/OCR4all/LAREX.git
• Compile: mvn clean install -f LAREX/pom.xml.
• Copy or link the created war file to tomcat:
  • Copy cp LAREX/target/Larex.war /var/lib/tomcat8/webapps/Larex.war
  • Link: sudo ln -s $PWD/LAREX/target/Larex.war /var/lib/tomcat8/webapps/Larex.war
• Start Tomcat: systemctl start tomcat8
  • (Restart Tomcat via systemctl restart tomcat8)
  • (To start Tomcat automatically at system boot systemctl enable tomcat8)

This guide uses eclipse to simplify the setup on Windows

• Install Eclipse IDE for Enterprise Java Developers from the official website
• Download tomcat 8 or up from the official website
• Create a tomcat server in eclipse:
  • Open Window -> Show View -> Other... -> Server -> Servers
  • Click prompt to add a new server, select Apache -> Tomcat <version> Server -> Next-> set your Tomcat installation directory -> Finish.
• Clone Repository:
  • File -> Import -> Git -> Projects from Git -> Clone URI -> Set URI: https://github.com/OCR4all/LAREX.git -> [✓] master -> Next > -> Next > -> Import as gernal project -> Finish
• Set as Maven Project:
• Import Project
  • Right click on Larex -> Configure -> Convert to Maven Project -> Finish
• Update maven project (if not updated automatically)
  • Rightclick on Larex -> Maven -> Update Project... -> OK
• Start Larex
  • Right click on Larex -> Run As -> Run on Server.

This guid uses homebrew (please adjust accordingly for your setup)

• Install Homebrew (see https://brew.sh/).
• Install required packages:
  • brew cask install java
  • brew install tomcat git maven
• (Verify tomcat installation):
  • brew services list tomcat should be listed in the output of this command
• Clone Repository:
  • git clone https://github.com/OCR4all/LAREX.git
• Compile:
  • run mvn clean install -f LAREX/pom.xml.
• Copy or link the created war file to tomcat
  • Copy: cp LAREX/target/Larex.war /usr/local/Cellar/tomcat/[version]/libexec/webapps/Larex.war
  • Link: ln -s $PWD/LAREX/target/Larex.war /usr/local/Cellar/tomcat/[version]/libexec/webapps/Larex.war
• Start Tomcat:
  • brew services start tomcat
  • (Restart Tomcat via brew services restart tomcat)

Go to localhost:8080/Larex.

You can add your own books by copying them to src/webapp/resources/books

(Or an alternative direction set in the config file. See section Configuration for more information).

Book directories must have the following structure:

bookDir/
├── <book_name>/ 
│    ├── <page_name>.png 
│    └── <page_name>.xml
└── <book2_name>/
     └── …

Detailed information about the usage of LAREX can be found in the OCR4all getting started guides.

See sections and chapters about Segmentation, Ground Truth Correction and Post Correction.

Larex contains a configuration file (src/webapp/WEB-INF/larex.config) with a few settings that can be set before running the application.

The setting bookpath sets the file path of the books folder.

e.g. bookpath:/home/user/books (Linux)

e.g. bookpath:C:\Users\user\Documents\books (Windows)

Larex will load the books from this folder.

[default /src/main/webapp/resources/books]

The setting localsave tells the application how to handle results locally when saved.

<mode>=[bookpath|savedir|none]

bookpath: save the result in the bookpath

savedir: save the result in a defined savedir

none: do not save the result locally [default]

e.g. localsave:bookpath

The setting savedir is needed if localsave mode is set to "savedir".

e.g. savedir:/home/user/save (Linux)

e.g. savedir:C:\Users\user\Documents\save (Windows)

The setting websave tells the application how to handle results on the browser side when saved.

<value>=[true|false]

true: download the result after saving [default]

false: no action after saving

e.g. websave:true

Set the accessible modes in the Larex gui <value>=[[segment][edit][lines][text]] A combination of the modes "segment", "edit", "lines" and "text" can be set as a space separated string. e.g. modes:segment lines

The order of those modes in the string also determines which mode is opened on startup, with the first in the list being opened as main mode. The mode "segment" can be replaced with "edit" in order to hide all auto segmentation features. ("edit" will be ignored if both are present)

[Default]modes:segment lines text

This setting enables or disables the direct open feature.

<value>=[enable|disable]

This feature allows users to load a book from everywhere on the servers drive aswell as to alter the options websave, localsave and savedir.

enable: enable direct request

disable: disable direct request [default]

e.g. directrequest:enable

This feature should be used with caution but is very useful when using Larex in a workflow with other web applications. (e.g. in docker)

The easiest direct request would be via a html form with the values bookpath, bookname, websave (optional), localsave (optional) and savedir (optional).

<form action="http://localhost:8080/Larex/direct" method="POST">
	bookpath: <input type="text" name="bookpath"/><br>
	bookname: <input type="text" name="bookname"/><br>
	websave: <input type="text" name="websave"/><br>
	localsave: <input type="text" name="localsave"/><br>
	savedir: <input type="text" name="savedir"/><br>
	modes: <input type="text" name="modes"/><br>
	<input type="submit"/>
</form>

This setting enables or disables OCR4all UI mode.

<value>=[enable|disable]

This setting allows displaying and/or hiding certain UI elements when LAREX is used in combination with OCR4all.

enable: enable OCR4all UI mode

disable: disable OCR4all UI mode [default]

e.g. ocr4all:enable

Reul, Christian; Springmann, Uwe; Puppe, Frank: LAREX – A semi-automatic open-source Tool for Layout Analysis and Region Extraction on Early Printed Books. In Proceedings of the 2nd International Conference on Digital Access to Textual Cultural Heritage (2017). ACM. Draft available at arXiv.

Reul, Christian; Dittrich, Marco; Gruner, Martin: Case Study of a highly automated Layout Analysis and OCR of an incunabulum: ‘Der Heiligen Leben’ (1488). In Proceedings of the 2nd International Conference on Digital Access to Textual Cultural Heritage (2017). ACM. Draft available at arXiv.