Mathematical Knowledge Browser with Automatic Hyperlink Detection

摘要

Mathematical OCR (Optical Character Recognition) systems retrieve character sequences and the structure of mathematical formulae from raster images scanned from mathematical documents. In this paper a method for detecting hyperlinks, e.g. formula links, from mathematical OCR output is described. We also experimentally demonstrated the effectiveness of the method. By using the method we implemented a prototype system of a mathematical knowledge browser that helps people read mathematical articles.