Handwritten text recognition applied to the manuscript production of the Carthusian Monastery of Herne in the Fourteenth Century
This paper contributes to the assessment of various – possibly impacting – factors during the collection of ground truth data for training HTR-systems. By scrutinising different parameters (e.g. scribal hands, handwriting styles, spelling profiles, textual genres, etc.) we will report on the impact of various train-target combinations.
Verfasser: | |
---|---|
Dokumenttyp: | conferencePaper |
Erscheinungsdatum: | 2023 |
Schlagwörter: | Paper / Long Presentation / handwritten text recognition / digital manuscript studies / Middle Dutch literature / scribal profiling / digital libraries creation / management / and analysis / manuscripts description / representation / optical character recognition and handwriting recognition / scholarly editing and editions development / analysis / and methods / Book and print history / Humanities computing / Linguistics / Literary studies |
Sprache: | Englisch |
Permalink: | https://search.fid-benelux.de/Record/base-27466582 |
Datenquelle: | BASE; Originalkatalog |
Powered By: | BASE |
Link(s) : | https://zenodo.org/record/8107391 |