Abstract
While bi-directional long short-term (BLSTM) neural network have been demonstrated to perform very well for English or Arabic, the huge number of different output classes (characters) encountered in many Asian fonts, poses a severe challenge. In this work we investigate different encoding schemes of Bangla compound characters and compare the recognition accuracies. We propose to model complex characters not as unique symbols, which are represented by individual nodes in the output layer. Instead, we exploit the property of long-distance-dependent classification in BLSTM neural networks. We classify only basic strokes and use special nodes which react to semantic changes in the writing, i.e., distinguishing inter-character spaces from intra-character spaces. We show that our approach outperforms the common approaches to BLSTM neural network-based handwriting recognition considerably.
Chapter PDF
Similar content being viewed by others
References
Bhattacharya, N., Pal, U., Kimura, F.: A System for Bangla Online Handwritten Text. In: 12th Int’l Conf. on Document Analysis and Recognition, pp. 1367–1371 (2013)
Bhattacharya, U., Guin, K., Parui, S.K.: Direction Code Based Features for Recognition of Online Handwritten Characters of Bangla. In: 9th Int’l Conf. on Document Analysis and Recognition, vol. 1, pp. 58–62 (2007)
Bhattacharya, U., Nigam, A., Rawat, Y.S., Guin, K.: An Analytic Scheme for Online Handwritten Bangla Cursive Word Recognition. In: 11th Int’l Conf. Frontiers in Handwriting Recognition, pp. 320–325 (2008)
Fink, G., Vajda, S., Bhattacharya, U., Parui, S.K., Chaudhuri, B.B.: Online Bangla Word Recognition Using Sub-Stroke Level Features and Hidden Markov Models. In: Int’l Conf. of Frontiers in Handwriting Recognition, pp. 393–398 (2010)
Frinken, V., Bhattacharya, N., Pal, U.: Design of Unsupervised Feature Extraction System for On-Line Bangla Handwriting Recognition. In: 11th IAPR International Workshop on Document Analysis Systems (page accepted for publication, 2014)
Frinken, V., Peter, T., Fischer, A., Bunke, H., Do, T.-M.-T., Artieres, T.: Improved Handwriting Recognition by Combining Two Forms of Hidden Markov Models and a Recurrent Neural Network. In: Jiang, X., Petkov, N. (eds.) CAIP 2009. LNCS, vol. 5702, pp. 189–196. Springer, Heidelberg (2009)
Garai, G., Chaudhuri, B.B., Pal, U.: Online Handwritten Indian Script Recognition: A Human Motor Function Based Framework. In: 16th Int’l Conference on Pattern Recognition, vol. 3, pp. 164–167 (2002)
Gers, F., Schmidhuber, J.: Recurrent Nets that Time and Count. In: IEEE-INNS-ENNS Joint Conf. on Neural Networks, vol. 3, pp. 189–194 (2000)
Graves, A.: Offline Arabic Handwriting Recognition with Multidimensional Neural Networks. In: Guide to OCR for Arabic Scripts, pp. 297–314. Springer (2012)
Graves, A., Fernández, S., Gomez, F., Schmidhuber, J.: Connectionist Temporal Classification: Labelling Unsegmented Sequential Data with Recurrent Neural Networks. In: 23rd Int’l Conf. on Machine Learning, pp. 369–376 (2006)
Graves, A., Liwicki, M., Fernández, S., Bertolami, R., Bunke, H., Schmidhuber, J.: A Novel Connectionist System for Unconstrained Handwriting Recognition. IEEE Transaction on Pattern Analysis and Machine Intelligence 31(5), 855–868 (2009)
Graves, A., Schmidhuber, J.: Framewise Phoneme Classification with Bidirectional LSTM Networks. In: Int’l Joint Conf. on Neural Networks, vol. 4, pp. 2047–2052 (2005)
Katayama, Y., Uchida, S., Sakoe, H.: A new HMM for On-Line Character Recognition using Pen-Direction and Pen-Coordinate Features. In: 19th Int’l Conf. on Pattern Recognition, pp. 1–4 (2008)
Kittler, J., Hatef, M., Duin, R.P., Matas, J.: On Combining Classifiers. IEEE Trans. on Pattern Analysis and Machine Intelligence 20(3), 226–239 (1998)
Kuncheva, L.I.: Combining Pattern Classifiers: Methods and Algorithms. Wiley-Interscience (2004)
Mondal, T., Bhattacharya, U., Parui, S.K., Das, K., Mandalapu, D.: On-line Handwriting Recognition of Indian Scripts – The First Benchmark. In: Int’l Conf. of Frontiers in Handwriting Recognition, pp. 200–205 (2010)
Nakagawa, M., Tokuno, J., Zhu, B., Onuma, M., Oda, H., Kitadai, A.: Recent Results of Online Japanese Handwriting Recognition and its Applications. In: Doermann, D., Jaeger, S. (eds.) SACH 2006. LNCS, vol. 4768, pp. 170–195. Springer, Heidelberg (2008)
Parui, S.K., Bhattacharya, U., Chaudhuri, B.B.: Online Handwritten Bangla Character Recognition Using HMM. In: Int’l Conf. on Pattern Recognition, pp. 1–4 (2008)
Parui, S.K., Bhattacharya, U., Shaw, B., Guin, K.: A Hidden Markov Models for Recognition of Online Handwritten Bangla Numerals. In: 41st National Annual Convention of CSI, pp. 27–31 (2006)
Plamondon, R., Srihari, S.N.: On-Line and Off-Line Handwriting Recognition: A Comprehensive Survey. IEEE Transaction on Pattern Analysis and Machine Intelligence 22(1), 63–84 (2000)
Rabiner, L.: A Tutorial on Hidden Markov Models and Selected Applications in Speech Recognition. Proceedings of the IEEE 77(2), 257–286 (1989)
Roy, K., Sharma, N., Pal, T., Pal, U.: Online bangla handwriting recognition system. In: 6th Int’l Conf. on Advances in Pattern Recognition, pp. 117–122 (2007)
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2014 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Frinken, V., Bhattacharya, N., Uchida, S., Pal, U. (2014). Improved BLSTM Neural Networks for Recognition of On-Line Bangla Complex Words. In: Fränti, P., Brown, G., Loog, M., Escolano, F., Pelillo, M. (eds) Structural, Syntactic, and Statistical Pattern Recognition. S+SSPR 2014. Lecture Notes in Computer Science, vol 8621. Springer, Berlin, Heidelberg. https://doi.org/10.1007/978-3-662-44415-3_41
Download citation
DOI: https://doi.org/10.1007/978-3-662-44415-3_41
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-662-44414-6
Online ISBN: 978-3-662-44415-3
eBook Packages: Computer ScienceComputer Science (R0)