A Mobile System for Historical Manuscripts Capturing, Recognition and Classification

Abstract

This paper proposes a mobile system to capture, recognize and classify historical manuscripts automatically. The user can take a snapshot of a manuscript through his/her mobile camera and send to a dedicated server. An algorithm will then be applied to analyze the received image through features extractions in order to determine its estimated century. These results will be sent back to the user and displayed on his/her mobile screen. We have also developed another part on the mobile application which allows the user to upload a given manuscript along with its information like century, author name, and location that will be stored in a KERTAS dataset after verification and validation.