A Skew Resistant Method for Persian Text Segmentation

Using OCR programs is one of the best ways to convert written and printed documents into digital form. The first phase in OCR is segmenting the input image and identifying text and non-text regions. This paper proposes a new method for segmentation of Persian printed texts which is based on the ink spread effect. Considering that the Persian scripts are very different from the English script, most methods proposed for the English script have not rendered good results for the Persian scripts. The method proposed in this paper has been designed considering the special features of the Persian scripts. In addition, one of the most important characteristics of this method is resistance to skew. Moreover, the proposed approach is directly applicable to Arabic scripts