Ancient degraded document image binarization based on texture features

Abstract

In this paper, we present a promising method for binarization of historical and degraded document images, based on texture features. The proposed method is an adaptive threshold-based. This latter is computed by using a descriptor based on a co-occurrence matrix. The proposed method is tested objectively, using DIBCO dataset degraded documents and subjectively, using a set of ancient degraded documents provided by a national library. The results are satisfactory and promising, and present an improvement to classical methods.