Comparison of OCR Accuracy on Early Printed Books using the Open Source Engines Calamari and OCRopus.
Christoph WickChristian ReulFrank PuppePublished in: J. Lang. Technol. Comput. Linguistics (2018)
Keyphrases
- open source
- optical character recognition
- computational cost
- high accuracy
- prediction accuracy
- open source software
- source code
- post processing
- error rate
- precision and recall
- highly accurate
- scanned documents
- machine learning
- improved accuracy
- character recognition
- correlation coefficient
- high precision
- classification accuracy
- preprocessing
- genetic algorithm