Conversion of scientific texts with OCR or manually?

The SZS is currently co-operating with the University of Kyushu, Japan. There at the Faculty of Mathematics the system INFTY has been developed under the leadership of Professor Suzuki. This system provides a text identification software (OCR) which is able to identify mathematical texts and to convert them into various output formats. Since mathematical texts - comparable to notes in music - are described graphically, conventional OCR programmes used to be able to interpret mathematical texts as graphics only.

In a joint project which is being supervised by a blind woman student at the Faculty of Computer Science at the University of Karlsruhe, this system is going to be tested. Since at present texts converted by INFTY are still very much in need of subsequent manual corrections, this project is intended to evaluate whether it is more efficient to convert texts manually - as has been done in the past - or whether the use of INFTY presents any real help. Thus, the project is supposed to supply researchers and users alike with incentives and suggestions for improvement. By now, the INFTY system enables the output of texts in LaTeX and XML. The output of texts in the ASCII mathematical script (AMS), which is being used at the SZS, is being envisaged.

The possible result of these efforts will be an OCR software which allows the conversion of mathematical texts.

more: infty