Automatic Segmentation of Handwritten Arabic Documents.

This paper presents an automatic extraction of handwritten components. In the first hand we present two more used methods in the literature. The first method is based on the MM. The second method is based on HT. In the second hand we propose a hybrid method based on the combination between the HT and the MM. These methods are evaluated on CENPARMI-Arabic Checks Database, in order to extract the handwritten components existing in the check (numerical amount, literal amount and date zone). We present a concept for automatic evaluation of the results, based on a label tools for the different part of the used documents. We achieve a correct extraction rate of 96.5% for the proposed method. [ABSTRACT FROM AUTHOR]