Locating Text in Historical Collection Manuscripts

It is common that documents belonging to historical collections are poorly preserved and are prone to degradation processes. The aim of this work is to leverage state-of-the-art techniques in digital image binarization and text identification for digitized documents allowing further content exploitation in an efficient way. A novel methodology is proposed that leads to preservation of meaningful textual information in low quality historical documents. The method has been developed in the framework of the Hellenic GSRT-funded R&D project, D-SCRIBE, which aims at developing an integrated system for digitization and processing of old Hellenic manuscripts. After testing of the proposed method on numerous low quality historical manuscripts, it has turned out that our methodology performs better compared to current state-of-the-art adaptive thresholding techniques.