Abstract
In this paper, an efficient algorithm for automatic outline capture of digital character images is presented. This method is well suited for non-Roman languages like Arabic, Urdu, Persian, etc. Contemporary word processing systems store shapes of the characters in terms of their outlines, and outlines are expressed as cubic Bezier curves. The process of capturing outlines includes steps detection of boundary, discorvering Corner Points and Break Points and fitting the curve. The work done in this paper, fully automate the above process and produces optimal results.