Skew Estimation in Persian Documents: A Novel Approach

In this paper, a robust method is proposed for estimating skew angle of Persian document images. The proposed algorithm applies morphological closing on blotched image to obtain image with text line blocks. Later, the text line blocks are subjected to thinning to obtain the skeleton and convert the text into line like structures used to estimate the skew angle of document image. The results of the experiments emphasizes that the proposed algorithm works well irrespective of the type of script even for a wide range of skew angle (with in plusmn90deg)