Methodology for fast skew angle detection and correction by using the Wavelet and the Hough transform: Application for Arabic printed documents

摘要

The purpose of this work is to present a new methodology for fast detecting the skew angle and correction for Arabic printed document based on Wavelet Transform (WT) and Hough Transform (HT). The application of the Hough transform presents a good solution for skew angle detection; however, this approach requires an important memory space and high computing time. For this, we have suggested the use of the WT in order to reduce the number of points and the computing time. This technique is based principally on the high band frequency of the wavelet transform. The proposed methodology has been applied and tested for 100 different documents. Obtained results indicate that the suggested approach gave good performance, the accuracy result of the skew angle estimation is very good detected and corrected, In addition, the computing time is well decreases compared to the alternative approaches. The technique has been applied for Arabic printed document, but it can be applied for any documents.