Graph Based Keyword Spotting in Medieval Slavic Documents - A Project Outline

Abstract

The objective of the planned project is to adapt a recent graph matching framework developed by the applicant to the problem of keyword spotting. The overall question to be answered is whether or not graph based representation and especially graph matching techniques can be beneficially employed for keyword spotting. For testing this novel keyword spotting framework, the Miroslav Gospels will be used. Miroslav Gospels is a 362-page illuminated manuscript Gospel Book on parchment with very rich decorations, which was inscribed on UNESCO’s Memory of the World Register in recognition of its historical value. It is one of the oldest surviving documents written in Old Church Slavonic. We plan to make the extracted word graphs from the Miroslav Gospels publicly available for further developments in graph based keyword spotting.