Projects & Grants

Internal Grant Competition DGC





Optical Character Recognition in Old Prints
Project IdSGS03/ÚVAFM/20
Main solverdoc. Mgr. Radek Čech, Ph.D.
Periodr1/2020 - 12/2020
ProviderSpecifický VŠ výzkum
Statefinished
AnotationThe goal of the project is to create software for optical character recognition of old prints (from the 17th century). The detection and recognition of individual characters will be based on machine learning techniques, concretely on deep neural networks. This will allow us to extract the information needed for linguistic analysis (models of a rank-frequency distribution of types in individual prints, tests of linguistic hypotheses, etc.).