A Two-OCR Engine Method for Digitized Swedish Newspapers.
Dana DannéllsLars BjörkOve DirdalTorsten JohanssonPublished in: CLARIN Annual Conference (2020)
Keyphrases
- preprocessing
- experimental evaluation
- synthetic data
- data sets
- dynamic programming
- computational cost
- segmentation method
- high accuracy
- clustering method
- computational complexity
- pairwise
- high precision
- classification method
- detection algorithm
- optimization algorithm
- computationally efficient
- input data
- probabilistic model
- prior knowledge
- information retrieval