Page Segmentation of Persian/Arabic Printed Text Using Ink Spread Effect

Nowadays, OCR (optical character recognition) is widely used for converting written documents to digital documents. One of the OCR phases is page segmentation. In page segmentation, text regions must be found in input image. In addition, text parts like text columns must be separated. In this paper, a new method for segmenting Persian/Arabic printed text is proposed. This method is based on ink spread effect idea, a new idea that has particular features. Main features of Persian/Arabic scripts are considered in designing this method. This method is skew resistant and can segment text within frames and tables or regions with gray background