Comparison of Skewness Normalization Methods on Javanese Script Documents Input

The scan of a text document that is visually quite perpendicular to its axis is very important. Correction operations on image slope will make the image visually perpendicular. This paper will compare the method of tilt detection both moment-based, center of gravity, and hough transformation on a printed image of a modern Javanese text document. The results of this study will be very useful in the line segmentation of document images which are part of the transliteration of Javanese script text documents. The data used in this study is the result of scanning Javanese literary books entitled “Hamong Tani”. On each page of the scan, the slope angle of the coordinate axes is tested using the three selected slope detection methods. The next step is the line segmentation process using the vertical projection method to see the success of each method in making slope corrections. Based on the result of comparation shown that the Hough Transform methode work better more than two methods else.