OCR-D at the ICDAR 2019 in Sydney
The OCR-D paper “okralact - a multi-engine Open Source OCR training system (Konstantin Baierer, Rui Dong, Clemens Neudecker) was accepted for the 5th International Workshop on Historical Document Imaging and Processing HIP 2019 (https://www.primaresearch.org/hip2019/) in the context of the ICDAR Conference 2019 in Sydney (https://icdar2019.org/).
The source code for the prototype for a multi-engine OCR Training infrastructure presented there is available at https://github.com/Doreenruirui/okralact. Training infrastructure is available at https://github.com/Doreenruirui/okralact for find. With “Dataset of Pages from Early Printed Books with Multiple Font Groups” the module project for the development of a model repository and an automatic font recognition for OCR-D is also at the same workshop with a paper of Mathias Seuret, Saskia Limbach, Nikolaus Weichselbaumer, Andreas Maier and Vincent Christlein.