Patent
1995-05-05
1997-10-14
Zele, Krista M.
395 26, 395 261, 395 279, G10L 500
Patent
active
056779909
ABSTRACT:
A multipass recognition strategy selects the N-best hypotheses resulting from each pass and propagates these N-best to the next pass. This strategy outperforms conventional hidden Markov model recognizers using a grammar constraining all possible names. Real time recognition of continuously spelled names is made feasible, in part, because the processor-intensive costly constraints are applied, if at all, in the 4th pass, after the system has produced a much smaller dynamic grammar.
REFERENCES:
patent: 4742547 (1988-05-01), Watanabe
patent: 4748670 (1988-05-01), Bahl et al.
patent: 4831550 (1989-05-01), Katz
patent: 5241619 (1993-08-01), Schwartz et al.
patent: 5268990 (1993-12-01), Cohen et al.
patent: 5454063 (1995-09-01), Rossides
patent: 5467425 (1995-11-01), Lau et al.
patent: 5502791 (1996-03-01), Nishimura et al.
"The N-Best Algorithm: An Efficient and Exact Procedure for Finding the N Most Likely Sentence Hypotheses", R. Schwartz et al., IEEE, pp. 81-84 Dec. 1990.
"A Tree Trellis Based Fat Search for Finding the N Best Sentence Hypotheses in Continuous Speech Recogntion", F.K. Soong et al., IEEE, pp. 705-708 Dec. 1991.
"An Algorithm of High Resolution and Efficient Multiple String Hypothesization for Continuous Speech Recognition Using Inter-Word Models", W. Chou et al., IEEE, pp. II-153 to II-156 Dec. 1994.
Schwarz, R. and Austin, S., Efficient, High-Performance Algorithms for N-Best Search, BBN Systems and Technologies Inc., 10 Moulton St., Cambridge, MA 02138, pp. 6-11 Jun. 1990.
Cole, R.; Roginski, K.; Fanty, M., English Alphabet Recognition With Telephone Speech, Oregon Graduate Institute of Science and Technology, 19600 N.W. Von Neumann Dr., Beaverton, OR 97006, pp. 479-482 date unknown.
Hermansky, H. et al., Compensation for the Effect of the Communication Channel in Auditory-Like Analysis of Speech (RASTA-PLP), U.S. West Advanced Technologies, 4001 Discovery Drive, Boulder, CO 80303, pp. 1367-1370 date unknown.
Damashek, M. "Gauging Similarity With n-Grams: Language-Independent Categorization of Text," Science, vol. 267, Feb. 10, 1995, pp. 843-848.
Mari, J. And Haton, J. "Automatic Word Recognition Based on Second-Order Hidden Markov Models," ICSLP 94, Yokohama, CRIN/CNRS & INRIA-Lorraine, BP 239 54506 Vandoeuvre-Les Nancy, France, pp. 247-250 (S07-16.1:S07-16.4).
Murveit, H. et al. Reduced Channel Dependence for Speech Recognition, SRI International, Speech Research & Technology Program, Menlo Park, CA 94025, pp. 280-284 date unknown.
Cole R. et al., "Speaker-Independent Name Retrieval From Spellings Using a Database of 50,000 Names," IEEE 1991, S5.19, pp. 325-328 Dec. 1994.
Junqua, J. et al., An N-Best Strategy, Dynamic Grammars and Selectively Trained Neural Networks for Real-Time Recognition of Continuously Speled Names Over the Telephone, Speech Technology Laboratory, Panasonic Technologies, Inc. 3888 State Street, Santa Barbara, CA 93105, CRIN-CNRS & INRIA Lorraine, BP 239-F54506 Vandoeuvre Les nancy, france; Institut Eurecom, BP 193-06904 Sophia Antipolis cedex, france May 12, 1995.
HTK: Hidden Markov Model Toolkit V1.4 Installation Guide, User Manual, Reference Manual, Programmer Manual, Cambridge University Engineering Department, Speech Group, P.C. Woodland, Oct. 2, 1992.
Naadeu, C. et al. "Filtering of Spectral Parameters for Speech Recognition," ICSLP 94, Yokohama, AT&T Bell Laboratories, Murray Hill, NJ 07974, pp. 1927-1930 (S31-24.1:S31-24.4) Dec. 1994.
Panasonic Technologies Inc.
Weaver Scott L.
Zele Krista M.
LandOfFree
System and method using N-best strategy for real time recognitio does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with System and method using N-best strategy for real time recognitio, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and System and method using N-best strategy for real time recognitio will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-1560493