This paper shows an approach for converting bitmap images of text glyphs into a vector format which is suitable for being embedded in XML representations of digitized documents. The focus is on a contour based vectorization method as the output can be easily transformed into SVG glyph descriptions. A concrete implementation is described and the results are discussed with special regard to the visual quality. The work is related to the development of a system for processing documents which are not suitable for current OCR methods. This is especially important in the field of retrospective digitization of historical works.
Stefan Pletschacher, Marcel Eckert, Arved C. H&uum