An efficient method for the skew normalization of a document image

An efficient method to normalize the skew of document images is proposed. The method is based on the idea that the baseline of a line of text is a straight line, and the inclination of this line with respect to the horizon is the skew angle. The detecting process is conducted in three stages. First, the lowest points of the symbols on a text line are extracted. Then, the height attribute of these symbols is used to filter out some unnecessary base points belonging to these symbols. Finally, a statistical approach is used to get the skew angle by using the position information of these points