Automatic Recognition of Serbian Medieval Manuscripts by applying the Transkribus software platform: current state and future perspectives

Abstract

The paper investigates the potentials of applying the model for the automatic recognition of (Russian) Church Slavonic manuscripts to Serbian medieval manuscripts written in various types of Cyrillic scripts by employing the Transkribus software platform. The analysis has shown: (a) that the use of the existing generic model for the recognition of Church Slavonic manuscripts can yield rather good results when applied to Serbian medieval manuscripts written in uncial or semiuncial script, (b) that the manuscripts written in the cursive script require the creation of a separate model, and (c) that the creation of a generic model within the Transkribus platform for the Serbian medieval manuscripts would make the process of digitization substantially faster, which in turn would lead to faster realization of tasks within the existing projects related to Serbian historical corpus linguistics and lexicography .