期刊名称：International Journal of Computer Science and Network
印刷版ISSN：2277-5420
出版年度：2014
卷号：3
期号：4
页码：220-224
出版社：IJCSN publisher
摘要：This work presents the preliminary results for identifying the ancient handwritten writing styles in Arabic manuscripts. The aim of this work is to discriminate between different writing styles appear in different dated ancient Arabic manuscripts in three Islamic historical ages (Contemporary, Ottoman and Mamluk) with objective to evaluate the capability of applying Zipf law on texture structures extracted by Spatial Gray Level dependency (SGLD) method, our approach relies on applying SGLD on handwritten ancient text image to extract the texture structures then applying Zipf’s distribution on document image texture structure. Based on this method we extract a feature presents the document image structure distribution and evaluate its efficiency for writing style identification
