Hybrid Training Data for Historical Text OCR

Kód výsledku v IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F49777513%3A23520%2F19%3A43958225" target="_blank" >RIV/49777513:23520/19:43958225 - isvavai.cz</a>

Výsledek na webu

<a href="http://dx.doi.org/10.1109/ICDAR.2019.00096" target="_blank" >http://dx.doi.org/10.1109/ICDAR.2019.00096</a>

DOI - Digital Object Identifier

<a href="http://dx.doi.org/10.1109/ICDAR.2019.00096" target="_blank" >10.1109/ICDAR.2019.00096</a>

Jazyk výsledku

angličtina

Název v původním jazyce

Hybrid Training Data for Historical Text OCR

Popis výsledku v původním jazyce

Current optical character recognition (OCR) systems commonly make use of recurrent neural networks (RNN) that process whole text lines. Such systems avoid the task of character segmentation necessary for character-based approaches. A disadvantage of this approach is a need of a large amount of annotated data. This can be solved by using generated synthetic data instead of costly manually annotated ones. Unfortunately, such data is often not suitable for historical documents particularly for quality reasons. This work presents a hybrid approach for generating annotated data for OCR at a low cost. We first collect a small dataset of isolated characters from historical document images. Then, we generate historical looking text lines from the generated characters. Another contribution lies in the design and implementation of an OCR system based on a convolutional-LSTM network. We first pre-train this system on hybrid data. Afterwards, the network is fine-tuned with real printed text lines. We demonstrate that this training strategy is efficient for obtaining state-of-theart results. We also show that the score of the proposed systém is comparable or even better in comparison to several state-ofthe-art systems.

Název v anglickém jazyce

Hybrid Training Data for Historical Text OCR

Popis výsledku anglicky

Current optical character recognition (OCR) systems commonly make use of recurrent neural networks (RNN) that process whole text lines. Such systems avoid the task of character segmentation necessary for character-based approaches. A disadvantage of this approach is a need of a large amount of annotated data. This can be solved by using generated synthetic data instead of costly manually annotated ones. Unfortunately, such data is often not suitable for historical documents particularly for quality reasons. This work presents a hybrid approach for generating annotated data for OCR at a low cost. We first collect a small dataset of isolated characters from historical document images. Then, we generate historical looking text lines from the generated characters. Another contribution lies in the design and implementation of an OCR system based on a convolutional-LSTM network. We first pre-train this system on hybrid data. Afterwards, the network is fine-tuned with real printed text lines. We demonstrate that this training strategy is efficient for obtaining state-of-theart results. We also show that the score of the proposed systém is comparable or even better in comparison to several state-ofthe-art systems.

Druh

D - Stať ve sborníku

CEP obor

—

OECD FORD obor

10201 - Computer sciences, information science, bioinformathics (hardware development to be 2.2, social aspect to be 5.8)

Projekt

<a href="/cs/project/EF17_048%2F0007267" target="_blank" >EF17_048/0007267: VaV inteligentních komponent pokročilých technologií pro plzeňskou metropolitní oblast</a><br>

Návaznosti

P - Projekt vyzkumu a vyvoje financovany z verejnych zdroju (s odkazem do CEP)<br>S - Specificky vyzkum na vysokych skolach<br>I - Institucionalni podpora na dlouhodoby koncepcni rozvoj vyzkumne organizace

Rok uplatnění

2019

Kód důvěrnosti údajů

S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů

Název statě ve sborníku

The 15th IAPR International Conference on Document Analysis and Recognition

ISBN

978-1-72813-014-9

ISSN

1520-5363

e-ISSN

—

Počet stran výsledku

6

Strana od-do

565-570

Název nakladatele

IEEE

Místo vydání

Piscataway

Místo konání akce

Sydney

Datum konání akce

20. 9. 2019

Typ akce podle státní příslušnosti

WRD - Celosvětová akce

Kód UT WoS článku

—