Abstract
A Speech and Character Combined Recognition Engine (SCCRE) is developed for working on Personal Digital Assistants (PDA) or on mobile devices. In SCCRE, feature extraction from speech and character is carried out separately, but recognition is performed in an engine. The recognition engine employs essentially CHMM (Continuous Hidden Markov Model) structure and this CHMM consists of variable parameter topology in order to minimize the number of model parameters and reduce recognition time. This model also adopts our proposed SSMS (Successive State and Mixture Splitting) for generating context independent model. SSMS optimizes the number of mixtures through splitting in mixture domain and the number of states through splitting in time domain. When we applied our developed engine which adopts SSMS to speech recognition for mobile devices, SSMS can reduce total number of Gaussian up to 40.0% compared with the fixed parameter models at the same recognition performance. This leads that SSMS can reduce the size of memory for models to 65% and that for processing to 82%. Moreover, recognition time decreases 17% with SSMS model but still maintains the recognition rate.
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
Suk, S.Y., Kim, M.J., Chung, H.Y.: An on-line speech and character combined recognition system for multimodal interfaces. In: EALPIIT Proc., pp. 89–92 (2002)
Sin, B.K., Kim, J.: A Statistical Approach with HMMs for On-line Cursive Hangul (Korean Script) Recognition. In: Second International Conference on Document Analysis and Recognition Proc., Zuchuba, Japan, pp. 147–150 (1993)
Tong, H.: Determination of the order of a markov chain by Akaike‘s information criterion. Journal of Applied Probability 12, 488–497 (1975)
Li, D., Biem, A., Subrahmonia, J.: Hmm topology optimization for handwriting recognition. In: ICASSP Proc. (2001)
Takami, J., Sagayama, S.: A successive state splitting algorithm for efficient allophone modeling. In: ICASSP 1992 Proc., vol. 1, pp. 573–576 (1992)
Takaki, H., Mashahru, K., Akinori, I., Masaki, K.: A Study on HM-Nets using Decision Tree-based Successive Splitting. In: ICSP 1997 Proc., pp. 383–387 (1997)
Nakagawa, S.: A connected spoken word recognition method by O(n) dynamic programming pattern matching algorithm. In: ICASSP Proc., pp. 296–299 (1983)
Ralph, G., Stefan, M., Alex, W.: Run-on recognition in an on-line handwriting recognition system. Carnegie Mellon Univ. Press (1997)
Biem, A., Ha, J.Y., Subrahmonia, J.: Bayesian model selection criterion For HMM Topology Optimization. In: ICASSP Proc., pp. 989–992 (2002)
Jitsuhiro, T., Nakamura, S.: Automatic generation of non-uniform HMM structures based on variational Bayesian approach. In: ICASSP Proc., vol. 1, pp. 805–808 (2004)
Deng, L.: Distributed speech processing in MiPad’s multimodal user interface. IEEE Trans. Speech and Audio 10(8), 605–619 (2002)
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2006 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Kim, MJ., Suk, SY., Jung, HY., Chung, HY. (2006). A Speech and Character Combined Recognition Engine for Mobile Devices. In: Sha, E., Han, SK., Xu, CZ., Kim, MH., Yang, L.T., Xiao, B. (eds) Embedded and Ubiquitous Computing. EUC 2006. Lecture Notes in Computer Science, vol 4096. Springer, Berlin, Heidelberg. https://doi.org/10.1007/11802167_56
Download citation
DOI: https://doi.org/10.1007/11802167_56
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-540-36679-9
Online ISBN: 978-3-540-36681-2
eBook Packages: Computer ScienceComputer Science (R0)