Efficient OCR post-processing combining language, hypothesis and error models

Abstract

In this paper, an OCR post-processing method that combines a language model, OCR hypothesis information and an error model is proposed. The approach can be seen as a flexible and efficient way to perform Stochastic Error-Correcting Language Modeling. We use Weighted Finite-State Transducers (WFSTs) to represent the language model, the complete set of OCR hypotheses interpreted as a sequence of vectors of a posteriori class probabilities, and an error model with symbol substitutions, insertions and deletions. This approach combines the practical advantages of a de-coupled (OCR + post-processor) model with the error-recovery power of a integrated model. © 2010 Springer-Verlag Berlin Heidelberg.