Computer-assisted Ancient Documents Re-organization

Abstract

A large number of uninterpreted ancient documents exist in Asia, which record a large amount of information about the politics, economy, and culture. Since the beginning of oracle bone inscription, characters have been developing and evolving, and the forms of ancient documents have also been changing. As a large amount of information is preserved in ancient documents, the interpretation and collation of ancient documents are of great importance. However, due to the age, many ancient documents are moth-eaten, stained, damaged, and other problems. And most ancient documents are written in a style that is no longer in use, leaving few experts able to interpret them. This paper from oracle bone inscription, ancient hand-written character, rubbings three diferent types of ancient documents, introduces the recognition technology of assisted ancient documents re-organization including image processing, deep learning, and data analysis and deep learning combined method.