Serbian Early Printed Books: Towards Generic Model for Automatic Text Recognition using Transkribus

Abstract

The paper describes the process of creating and evaluating a new version of the generic model for automatic text recognition of Serbian Church Slavonic printed books within the Transkribus software platform, based on the principles of artificial intelligence and machine learning. The generic model Dionisio 2.0. was created on the materials of Serbian Church Slavonic books from various printing houses of the 15th and 16thcenturies (Cetinje, Venice, Goražde, Mileševa, Gračanica, Belgrade and Mrkša’s Church), and, during the evaluation of its performance, it was noticed that CER was about 2–3%. The Dionisio 2.0. model will be publicly available to all users of the Transkribus software platform in the near future.