Image enhancement of old manuscripts using machine learning

Abstract

Old manuscripts are important documents aging up to 1000 years discovered from all over the world and suffer from reimbursements and rottenness. These damages may be substantial making it problematical for researchers to scrutinize it. These damages contain varying contrast changes, degraded quality of the paper over time due to storage conditions, material of the document itself etc. To make these documents legible we consider image enhancement methods to enhance the quality standard of the manuscripts and make them constructive. This paper uses the concepts of machine learning algorithms in image enhancement and detection of patterns. In this paper, Ratio Rule machine learning algorithm is used out for various algorithms such as Bayesian Nets, Decision Trees and Nearest Neighbors algorithms.