Transkribus - A Service Platform for Transcription, Recognition and Retrieval of Historical Documents

Abstract

Transkribus is a comprehensive platform for the computer-aided transcription, recognition and retrieval of digitized historical documents. The main user interface is provided via an open-source desktop application that incorporates means to segment document images, to add a transcription and to tag entities within. The desktop application is able to connect to the platform's backend, which implements a document management system as well as several tools for document image analysis, such as layout analysis or automatic/handwritten text recognition (ATR/HTR). Access to documents, uploaded to the platform, may be granted to other users in order to collaborate on the transcription and to share results.