MAYASTROUN: A Multilanguage Handwriting Database
Sourour Njah, Badreddine Ben Nouma, Hala Bezine, Adel M. Alimi · 2012
To test the performance of handwriting on-line or offline recognition system, to measure and experimenting handwriting, to valid psychologist's laws, to compare and to make competitions between existing handwriting system, we must use databases. In this paper, we present the new dual handwriting database: "MAYASTROUN". The developed database contains a large lexicon of unconstrained cursive Arabic and Latin texts, words, characters, digits, mathematical expressions and signatures. The collected data contains more than 67825 and written by 355 writers. To facilitate experiments and research, this database offers four different extensions of each file and interactive software. The database architecture is described in details in this paper. The MAYASTROUN-database is available for the purpose to improve handwriting research filed.