HANDWRITTEN TEXT RECOGNITION FOR CROATIAN GLAGOLITIC.
The paper presents and discusses recent advances in Handwritten Text Recognition (HTR) technologies for handwritten and early printed texts in Croatian Glagolitic script. After elaborating on the general principles of training HTR models with respect to the Transkribus platform used for these experi...
| Publicado en: | Slovo (05836255) no. 72; pp. 181 - 193 |
|---|---|
| Autor principal: | |
| Formato: | Artículo |
| Publicado: |
Slovo
2022
|
| Materias: | |
| Acceso en línea: | Ver este registro en EBSCOhost |
| Sumario: | The paper presents and discusses recent advances in Handwritten Text Recognition (HTR) technologies for handwritten and early printed texts in Croatian Glagolitic script. After elaborating on the general principles of training HTR models with respect to the Transkribus platform used for these experiments, the characteristics of the models trained are discussed. Specifically, the models use the Latin script to transcribe the Glagolitic source. In doing so, they transcribe ligatures and resolve abbreviations correctly in the majority of cases. The computed error rate of the models is below 6%, real-world performance seems to be similar. Using the models for pre-transcription can save a great amount of time when editing manuscripts and, thanks to fuzzy search (keyword spotting), even uncorrected HTR transcriptions can be used for various kinds of analysis. The models are publicly available via the Transkribus platform. Every scholar working on Glagolitic manuscripts and early printings is encouraged to use them. U radu se predstavljaju nedavni pomaci u tehnologiji prepoznavanja rukopisnoga teksta (HTR) namijenjenoj hrvatskoglagoljskim rukopisnim i ranim tiskanim knjigama. Nakon opisivanja općih načela strojne obuke HTR modela, iznose se značajke načela strojnoga učenja u platformi Transkribus, pogotovo modeli korištenja latinice u preslovljavanju glagoljskih tekstova. Pri tome se u većini slučajeva ispravno preslovljavaju ligature i razrješuju kratice. Dobivena čestota pogrešaka je manja od 6%, poput uobičajene čestote pogrešaka kada preslovljavanje provode stručne osobe. Primjena HTR modela u prvom stadiju preslovljavanja može uštedjeti puno vremena pri pripemi i uređivanju rukopisa za objavu, zahvaljujući pretraživanju (pretrazi po ključnim riječima), pa čak i neispravno HTR preslovljavanje može biti korišteno za različite raščlambe. Modeli su javno dostupni posredstvom platforme Transkribus. Potičemo sve znanstvenike koji obrađuju glagoljske rukopise i rane tiskane knjige da se njima koriste. |
|---|