Character-Level Alignment Using WFST and LSTM for Post-processing in Multi-script Recognition Systems - A Comparative Study
In this paper, two new techniques to correct the OCR errors are proposed, recurrent neural networks with Long-Short Term Memory (LSTM), and Weighted Finite State Transducers (WFSTs) with context-dependent confusion rules. Both methods are applied on OCR results of Latin, and Urdu Script. Especially Urdu script is very challenging to OCR. For building an error model using context-dependent confusion rules, the OCR confusions which appear in the recognition outputs are translated into edit operations using Levenshtein edit distance algorithm. The new LSTM model avoids the calculations that occur in searching the language model and it also makes the language model eligible to correct unseen incorrect words. Our generic approaches are language independent. The proposed supervised LSTM model is compared with the context-dependent error model and state-of-the-art single rule-based methods. The evaluation on Latin script shows the error rate of LSTM is 0.48 %, error model is 0.68 % and the rule-based model is 1.0 %. The evaluation shows that the accuracy of LSTM model on the Urdu testset is 1.58 %, while the accuracy of the error model is 3.8 % and OCR recognition results is 6.9 % for Urdu testset. LSTM showed best performance on both Latin and Urdu script. As such, experiments show that LSTM performs very well in language techniques, especially, post-processing.
Unable to display preview. Download preview PDF.
- 1.Al-Azawi, M., Afzal, M.Z., Breuel, T.M.: Normalizing historical orthography for OCR historical documents using LSTM. In: Proc. of the 2nd International Workshop on Historical Document Imaging and Processing, HIP 2013, pp. 80–85. ACM, New York (2013)Google Scholar
- 2.Al-Azawi, M.I.A., Liwicki, M., Breuel, T.M.: WFST-based ground truth alignment for difficult historical documents with text modification and layout variations. In: DRR Proc. SPIE (2013)Google Scholar
- 4.Frinken, V., Zamora-Martinez, F., Espana-Boquera, S., Castro-Bleda, M., Fischer, A., Bunke, H.: Long-short term memory neural networks language modeling for handwriting recognition. In: 21st ICPR, pp. 701–704 (November 2012)Google Scholar
- 6.Hassan, A., Noeman, S., Hassan, H.: Language independent text correction using finite state automata. In: International Joint Conference on NLP (2008)Google Scholar
- 8.Llobet, R., Navarro-Cerdan, J.R., Perez-Cortes, J.C., Arlandis, J.: Efficient OCR post-processing combining language, hypothesis and error models. In: Hancock, E.R., Wilson, R.C., Windeatt, T., Ulusoy, I., Escolano, F. (eds.) SSPR & SPR 2010. LNCS, vol. 6218, pp. 728–737. Springer, Heidelberg (2010)CrossRefGoogle Scholar
- 10.Mikolov, T., Deoras, A., Kombrink, S., Burget, L., Cernocky, J.: Empirical evaluation and combination of advanced language modeling techniques. In: Proc. of Inter. Speech Communication Association, Florence, Italy (2011)Google Scholar
- 11.Ul-Hasan, A., Bin Ahmed, S., Rashid, F., Shafait, F., Breuel, T.: Offline printed urdu nastaleeq script recognition with bidirectional LSTM networks. In: 12th Intern. Conf. on Document Analysis and Recognition, pp. 1061–1065 (2013)Google Scholar