A novel technique for estimation of skew in binary text document images based on linear regression analysis

Abstract

When a document is scanned either mechanically or manually for digitization, it often suffers from some degree of skew or tilt. Skew-angle detection plays an important role in the field of document analysis systems and OCR in achieving the expected accuracy. In this paper, we consider skew estimation of Roman script. The method uses the boundary growing approach to extract the lowermost and uppermost coordinates of pixels of characters of text lines present in the document, which can be subjected to linear regression analysis (LRA) to determine the skew angle of a skewed document. Further, the proposed technique works fine for scaled text binary documents also. The technique works based on the assumption that the space between the text lines is greater than the space between the words and characters. Finally, in order to evaluate the performance of the proposed methodology we compare the experimental results with those of well-known existing methods.

The Evaluation of the Initial Skew Rate for Printed Text

“…Another simple method based on boundary growing approach exclusively for printed text is proposed in [10]. Although simple method, making its refinement caused it to be time consuming.…”

“…Modification of the method [10] is presented in [11]. Proposed method advanced original method and solve some deficiencies.…”

“…Hence, those methods are in the scope of the paper. Hence, in this paper modification of the method based on the rectangular hull and its gravity center points [10] is presented. Initial method is extended by simplifying it.…”

“…This way, the morphological operation eliminates noise made by isolated black pixels in the image. Original algorithm main constraint [10] was assumption that text line spacing is always sufficient. Thus, boundary-growing region made by algorithm that extends text areas will not connect nearby text lines.…”

“…Main difference between original algorithm [10] and our approach are in text segmentation approach. In our case, current rectangular hull is not growing in all directions to the nearby rectangular hull of the next character.…”