Recognition of Mathematical Texts

Warning

This publication doesn't include Faculty of Arts. It includes Faculty of Informatics. Official publication website can be found on muni.cz.

Authors	HRDINA Miroslav SOJKA Petr
Year of publication	2013
MU Faculty or unit	Faculty of Informatics
Citation
Description	This work aims to resolve the OCR issue for EuDML project with the most important part being the recognition of mathematical content. Goal of this work was to create workflow for mathematical OCR, which transforms the mathematical publications to indexable recognized output. This was achieved by using the InftyReader OCR software. The chosen recognition workflow was tested and results were evaluated. During recognition numerous problems with the software and recognized data were encountered. Part of these problems were resolved, but some are still pending and will be resolved in the future.
Related projects:	The European Digital Mathematics Library Zastoupení ČR v European Research Consortium for Informatics and Mathematics