التعديل الذاتي لصور الوثائق المائلة

Abstract

A considerable number of computer researchers and users experience image
related problems pertaining the skews of the scanned documents which
result in irregular and improper document images compairing to the original
text images. Hence, the researcher attempts in the present study to overcome
this kind of problems through detecting and adjusting the skew angle of the text image to have the proper image. A number of algorithms are used to
calculate the skew angle in the printed text images. In the present study, an
Arabic printed text is used and introduced into the computer through
scanners in a form that makes the introduced text images deviaut of the
horizon. The skew angle is calculated by cutting the upper and lower empty
lines together with the right and left ones of the printed text. Then, the first
paragraph of the image taken is extracted after calculating its histogram
scale. The following step involves introducing the resultant part into Radon
Transform to produce an angle upon the value of which the rotation of the
image is based to have the proper text image. The algorithm applied to the
image of document texts having varying skew degrees yields good results as
far as accuracy is concerned reaching when the document is restored.