Skew detection and reconstruction based on maximization of variance of transition-counts

The input document images with skew can be a serious problem in the optical character recognition system. A robust method is proposed in this paper fo r skew detection and reconstruction in document images which can contain le ss text areas, high noises, tables, figures, flow-chart, and photos. The ba sic idea of our approach is the maximization of variance of transition coun ts for the skew detection and text-orientation determination. Once the skew angle is determined, the scanning-line model is applied to reconstruct the skew images. 103 documents with great varieties have been tested and succe ssfully processed. The average detection time of A4 size image is 4.56 s an d the reconstruction time is 5.52 s. The proposed approach is also compared with the existing algorithms published in the literature and our method ge ts some significant improvements in skew detection and reconstruction. (C) 1999 Pattern Recognition Society. Published by Elsevier Science Ltd. All ri ghts reserved.