Efficient OCR Post-Processing Combining Language, Hypothesis and Error Models

  • Rafael Llobet
  • J. Ramon Navarro-Cerdan
  • Juan-Carlos Perez-Cortes
  • Joaquim Arlandis
Part of the Lecture Notes in Computer Science book series (LNCS, volume 6218)


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.


Error Model Language Model Posteriori Probability Hypothesis Model Probable Path 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.


  1. 1.
    Mohri, M., Pereira, F., Riley, M.: The design principles of a weighted finite-state transducer library. Theoretical Computer Science 231, 17–32 (2000)zbMATHCrossRefMathSciNetGoogle Scholar
  2. 2.
    Allauzen, C., Riley, M., Schalkwyk, J., Skut, W., Mohri, M.: OpenFst: A General and Efficient Weighted Finite-State Transducer LIbrary. In: Holub, J., Žďárek, J. (eds.) CIAA 2007. LNCS, vol. 4783, pp. 11–23. Springer, Heidelberg (2007)CrossRefGoogle Scholar
  3. 3.
    Perez-Cortes, J.C., Amengual, J.C., Arlandis, J., Llobet, R.: Stochastic Error Correcting Parsing for OCR Post-processing. In: Proceedings of the ICPR, vol. 4, pp. 405–408 (2000)Google Scholar
  4. 4.
    Vidal, E., Thollard, F., de la Higuera, C., Casacuberta, F., Carrasco, R.: Probabilistic Finite-State Machines - Parts I and II. IEEE Trans. on Pattern Analysis and Machine Intelligence 27, 1013–1039 (2005)CrossRefGoogle Scholar
  5. 5.
    Garcia, P., Vidal, E.: Inference of k-testable languages in the strict sense and application to syntactic pattern recognition. IEEE Trans. on PAMI 12, 920–925 (1990)Google Scholar
  6. 6.
    Riley, M., Pereira, F., Mohri, M.: Transducer composition for context-dependent network expansion. In: Proc. of Eurospeech 1997 (1997)Google Scholar
  7. 7.
    Amengual, J., Vidal, E.: Efficient error-correcting viterbi parsing. IEEE Trans. on PAMI 20, 1109–1116 (1998)Google Scholar
  8. 8.
    Neuhoff, D.: The viterbi algorithm as an aid in text recognition. IEEE Trns. on Inf. Theory 21, 222–226 (1975)CrossRefGoogle Scholar
  9. 9.
    Berghel, H.L.: A logical framework for the correction of spelling errors in electronic documents. Information Processing and Management 23, 477–494 (1987)CrossRefGoogle Scholar
  10. 10.
    Hall, P., Dowling, G.: Approximate string matching. ACM Surveys 12, 381–402 (1980)CrossRefMathSciNetGoogle Scholar
  11. 11.
    Beaufort, R., Mancas-Thillou, C.: A Weighted Finite-State Framework for Correcting Errors in Natural Scene OCR. In: Proceedings of the Ninth International Conference on Document Analysis and Recognition, vol. 2, pp. 889–893 (2007)Google Scholar
  12. 12.
    Farooq, F., Jose, D., Govindaraju, V.: Phrase-based correction model for improving handwriting recognition accuracies. Pattern Recognition 42, 3271–3277 (2009)zbMATHCrossRefGoogle Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2010

Authors and Affiliations

  • Rafael Llobet
    • 1
  • J. Ramon Navarro-Cerdan
    • 1
  • Juan-Carlos Perez-Cortes
    • 1
  • Joaquim Arlandis
    • 1
  1. 1.Instituto Tecnologico de InformaticaUniversidad Politecnica de ValenciaValenciaSpain

Personalised recommendations