Preliminary Research on Computer-Assisted Transcription of Medieval Scripts in the Latin Alphabet using AI Computer Vision techniques and Machine Learning
A Romanian Exploratory Initiative
Keywords:Middle Ages, Latin writing, palaeography, Artificial Intelligence, Computer Vision, automatic transcription
The objective of the present paper is to introduce to a wider audience, at a very early stage of development, the initial results of a Romanian joint initiative of AI software engineers and palaeographers in an experimental project aiming to assist and improve the transcription effort of medieval texts with AI software solutions, uniquely designed and trained for the task. Our description will start by summarizing the previous attempts and the mixed-results achieved in e-palaeography so far, a continuously growing field of combined scholarship at an international level. The second part of the study describes the specific project, developed by Zetta Cloud, with the aim of demonstrating that, by applying state of the art AI Computer Vision algorithms, it is possible to automatically binarize and segment text images with the final scope of intelligently extracting the content from a sample set of medieval handwritten text pages.