Abstract
This paper presents a systematic investigation of the benefits resulting from the introduction of context-dependent modeling techniques in HMM-based large vocabulary handwriting recognition systems. It is shown how context-dependent units can be successfully introduced in complex handwriting systems by using so-called trigraphs, representing embedded characters within a word in its left and right context. With the introduction of suitable trigraphs we found relative error reduction rates up to 50% for writer-dependent recognition tasks with a 1000 word vocabulary. We could also verify these clear results for very large vocabularies with 30000 words, different writers and different unconstrained writing styles.
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Preview
Unable to display preview. Download preview PDF.
References
K. S. Nathan, J. Subrahmonia, and M. P. Perrone. Parameter Tying in Writer-Dependent Recognition of On-Line Handwriting. In Proc. IEEE Int. Conference on Pattern Recognition (ICPR), volume 3, pages 28–23, Vienna, 1996.
T. Starner, J. Makhoul, R. Schwartz, and G. Chou. On-Line Cursive Handwriting Recognition Using Speech Recognition Methods. In Proc. IEEE Int. Conference on Acoustics, Speech and Signal Processing (ICASSP), volume 5, pages 125–128, Adalaide, 1994.
J. R. Bellegarda, D. Nahamoo, and K. S. Nathan. Supervised Hidden Markov Modeling for On-Line Handwriting Recognition. In Proc. IEEE Int. Conference on Acoustics, Speech and Signal Processing (ICASSP), volume 5, pages 149–152, Adalaide, 1994.
G. Rigoll, A. Kosmala, J. Rottland, and Ch. Neukirchen. A Comparison between Continuous and Discrete Density Hidden Markov Models for Cursive Handwriting Recognition. In Proc. IEEE Int. Conference on Pattern Recognition, Vienna, 1996.
S. Manke, M. Finke, and A. Waibel. Combining Bitmaps with Dynamic Writing Information for On-Line Handwriting Recognition. In Proc. IEEE Int. Conference on Pattern Recognition (ICPR), pages 596–598, Jerusalem, 1994.
A. Kosmala, J. Rottland, and G. Rigoll. Improved On-Line Handwriting Recognition Using Context Dependent Hidden Markov Models. In Proc. Int. Conference on Document Analysis and Recognition, Ulm, 1997.
A. Kosmala, J. Rottland, and G. Rigoll. An Investigation of the Use of Trigraphs for Large Vocabulary Cursive Handwriting Recognition. In Proc. IEEE Int. Conference on Acoustics, Speech and Signal Processing (ICASSP), pages 3373–3376, Munich, 1997.
G. Rigoll. Maximum Mutual Information Neural Networks for Hybrid Connectionist-HMM Speech Recognition Systems. IEEE Transactions on Speech and Audio Processing, Special Issue on Neural Networks for Speech, 2 (1): 175–184, January 1994.
K. Lee. Context Dependent Phonetic Hidden Markov Models for Speaker Idependent Continuous Speech Recognition. IEEE Transactions on Acoustics, Speech and Signal Processing, 38 (4): 599–609, 1990.
S. J. Young and P. C. Woodland. The Use of State Tying in Continuous Speech Recognition. In Proc. Eurospeech, European Conference on Speech Communication and Technology, pages 2203–2206, Berlin, 1993.
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 1997 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Kosmala, A., Rottland, J., Rigoll, G. (1997). Large Vocabulary On-Line Handwriting Recognition with Context Dependent Hidden Markov Models. In: Paulus, E., Wahl, F.M. (eds) Mustererkennung 1997. Informatik aktuell. Springer, Berlin, Heidelberg. https://doi.org/10.1007/978-3-642-60893-3_26
Download citation
DOI: https://doi.org/10.1007/978-3-642-60893-3_26
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-540-63426-3
Online ISBN: 978-3-642-60893-3
eBook Packages: Springer Book Archive