Abstract
In this paper, we present a system for cursive script recognition. It is built upon several procedures for preprocessing, a neural network to recognize individual characters in a word, and a hidden Markov model for the recognition of complete words. As neural network we use a time delay neural network that receives input from a sliding window, which scans an input word from left to right. The network is trained with the back-propagation algorithm which is expanded by position invariant learning. Because of the enormous need of computation resources we developed a method to include more information in the learning procedure. As a result only about a third of the cycles in the training are needed to achieve the same recognition results. Alternatively keeping the number of training cycles constant, an increase in word recognition rate of about 5% to a total of 87% can be achieved.
Preview
Unable to display preview. Download preview PDF.
References
Y. S. Abu-Mostafa. Learning from Hints. Journal of Complexity, Vol. 10, No 1, pp. 165–178, 1994.
U. Bodenhausen, and S. Manke. Application oriented automatic structuring of time-delay neural networks for high performance character and speech recognition. ICANN-93, Proc. of the Int. Conf. on Artificial Neural Networks, Amsterdam, pp. 956–961, Springer Verlag, 1993.
T. M. Breuel. Recognition of Handwritten Responses on US Census Forms. In A. L. Spitz, A. Dengel, editors, Document Analysis Systems, World Scientific, pp. 237–264, 1995.
H. Bunke, M. Roth, and E.G. Schukat-Talamazzini. Off-line cursive handwriting recognition using hidden Markov models. In Pattern Recognition, Vol. 28, No. 9, pp. 1399–1413, 1995.
J.P. Dodel and R. Shinghal. Symbolic/neural recognition of cursive amounts on bank cheques. Proc. ICDAR'95, Montreal 1995. pp 15–18.
G. Kaufmann. H. Bunke and T.M. Ha. Recognition of cursively handwritten words using a combined normalization/perturbation approach. IWFHR 5, Sept. 2–5, 1996, University of Essex, England, pp 17–22.
J. D. Keeler, D.E. Rumelhart, and W.K. Leow. Integrated segmentation and recognition of hand-printed numerals. In David S. Touretzky Richard P. Lippmann, John E. Moody, editor, Neural Information Processing Systems 3, volume 3, pages 557–563. Morgan Kaufmann Publishers, 1989.
J. D. Keeler, D.E. Rumelhart, and W.K. Leow. Integrated segmentation and recognition of hand-printed numerals. MCC Technical Report ACT-NN-010-91, January 1991.
L.R. Rabiner. A tutorial on hidden markov models and selected applications in speech recognition. In Proceedings of the IEEE, volume 77, pages 257–286, February 1989.
H. Weissman, M. Schenkel, I. Guyon, C. Nohl. and D. Henderson. Recognition-Based Segmentation of On-line Run-on Handprinted Words: Input vs. Output Segmentation. Pattern Recognition. volume 27, number 3, pages 405–420, Pergamon Press, Oxford, 1994.
Author information
Authors and Affiliations
Editor information
Rights and permissions
Copyright information
© 1997 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Marti, UV., Kaufmann, G., Bunke, H. (1997). Cursive script recognition with time delay neural networks using learning hints. In: Gerstner, W., Germond, A., Hasler, M., Nicoud, JD. (eds) Artificial Neural Networks — ICANN'97. ICANN 1997. Lecture Notes in Computer Science, vol 1327. Springer, Berlin, Heidelberg. https://doi.org/10.1007/BFb0020279
Download citation
DOI: https://doi.org/10.1007/BFb0020279
Published:
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-540-63631-1
Online ISBN: 978-3-540-69620-9
eBook Packages: Springer Book Archive