IAM-HistDB

IAM-HistDB is a comprehensive research database was created that contains Latin texts from the 9^th century written in Carolingian script, Old German texts from the 13^th century written in Gothic script, and English texts from the 18^th century written in Longhand. The IAM-HistDB is one of the first publicly available research databases in the field.

Published online in 2012 it has already attracted over 100 users from international research institutes and has the potential to become a benchmark database. The creation of the database was supported by the research team of Prof. Dr. Micheal Stolz from the German Language Institute at the University of Bern.

IAM-HistDB contains handwritten historical manuscript images together with ground truth data for training and testing automatic handwriting recognition systems. It includes the three datasets

• Saint Gall Database - 9th century, Latin
• Parzival Database - 13th century, German
• Washington Database - 18th century, English

Sample pages of the IAM-HistDB datasets: Parzival, Saint Gall, George Washington

More Details on the HisDoc Project

• IAM-HistDB Link to the repository @ Uni Bern
• Module 1 - Document Image Analysis
• Module 2 - Handwriting Recognition
• Module 3 - Information Retrieval
• Integrated System - Full Processing Chain