Arabic Artistic Script Style Identification Using Texture Descriptors.
Zineb Kaoudja, Belal Khaldi, Mohammed Lamine Kherfi · 2020
Texture descriptors have been widely used for many computer vision tasks. Document Analysis (DA) is a hot topic of research, which aims at analyzing digitized documents containing different types of texts. In this paper, we design and evaluate a system for Arabic artistic style recognition using numerous texture descriptors. These descriptors have been chosen based on their effectiveness and the high performance they showed in other related works. To make it more comprehensive, the system has been evaluated using five different classifiers. Results indicated that texture descriptors appropriately fit the task and yielded almost perfect performances for most of the styles.