Line spectral frequencies and energy features in a robust signal recognition system
Patent 6009391 Issued on December 28, 1999. Estimated Expiration Date: August 6, 2017. Estimated Expiration Date is calculated based on simple USPTO term provisions. It does not account for terminal disclaimers, term adjustments, failure to pay maintenance fees, or other factors which might affect the term of a patent.
One embodiment of a speech recognition system is organized with speech input signal preprocessing and feature extraction followed by a fuzzy matrix quantizer (FMQ). Frames of the speech input signal are represented in a matrix by a vectorf of line spectral pair frequencies and energy coefficients and are fuzzy matrix quantized to respective vector f entries of a matrix codeword in a codebook of the FMQ. The energy coefficients include the original energy and the first and second derivatives of the original energy which increase recognition accuracy by, for example, being generally distinctive speech input signal parameters and providing noise signal suppression especially when the noise signal has a relatively constant energy over at least two time frame intervals. To reduce data while maintaining sufficient resolution, the energy coefficients may be normalized and logarithmically represented. A distance measure between f and f, d(f, f), is defined as ##EQU1## where the constants 댑, 댒, 댡 and 댢 are set to substantially minimize quantization error, ei is the error power spectrum of the speech input signal and a predicted speech input signal at the ith line spectral pair frequency of the speech input signal, the first G LSP frequencies are most likely to be frequency shifted by noise, and the last P+3 coefficients represent the three energy coefficients. This robust distance measure can be used to enhance speech recognition performance in generally any speech recognition system using line spectral pair based distance measures.
Other References
Rabiner, Lawrence and Juang, Biing-Hwang; "Fundamentals of Speech Recognition" 1993; pp. 190-195
Xydeas, C.S. Prof. and Cong, Lin "Robust Speech Recognition Using Fuzzy Matrix Quantisation, Neural Networks and Hidden Markov Models" Sep. 1996, pp. 1587-1590
Cong, Ling, Xydeas, Costas S. Prof. and Ferwood, Anthony F. Combining Fuzzy Vector Quantisation and Neural Network Classification for Robust Isolated Word Speech Recognition: Singapore ICCS 1994, pp. 884-887
Parsons, Thomas W,; "Voice and Speech Processing"; McGraw-Hill, Inc., New York, 1987; pp. 170-171
Xydeas, C.S. and Lin Cong; "Robust Speech Recognition Using Fuzzy Matrix Quantization and Neural Networks"; Proceedings of International Conference on Communication Technology; Beijing, China -ICCT '96; pp. 432-435; IEEE; New York (May 5-7,1996)
Cong, Lin; "A Study of Robust IWSR Systems"; PhD Thesis submitted to The University of Manchester School of Engineering, Division of Electrical Engineering; Manchester, United Kingdom; pp. 1-209, May 1996
Waibel, Alexander; "Neural Network Approaches for Speech Recognition"; Chapter 18 of Advances in Speech Signal Processing; edited by Sadaoki Furui and M. Mohan Sondhi; Marcel Dekker, Inc. ; New York, New York; 1992; pp. 555-595
Xydeas, C.S. and Cong, L.; "Combining Neural Network Classification with Fuzzy Vector Quantization and Hidden Markov Models for Robust Isolated Word Speech Recognition"; Signal Processing VIII Theories and Applications, vol. III; Proceedings of the IEEE International Symposium on Information Theory IEEE Press, 1995, p.174
Xydeas, C.S. and Cong, L; "Robust Speech Recognition in A Car Environment"; Presented at DSP95 International Conference on Digital Signal Processing, Jun. 26-28, 1995, Limassol, Cyprus; vol. 1, pp. 84-89
Cong, Lin, Prof. C.S. Xydeas, and Anthony Ferwood; "A Study of Robust Isolated Word Speech Recognition Based on Fuzzy Methods"; Presented at EUSIPCO-94, VII European Signal Processing Conference, Sep. 13-16, 1994; Scotland, UK.; 4 pages
Gibson, Jerry D.; "Coding, Transmission, and Storage"; Chapter 14, Speech Signal Processing, ofThe Electrical Engineering Handbook; Editor-in-Chief Richard C. Dorf; .COPYRGT.1993 by CRC Press, Inc.; pp. 279-314
Gersho, Allen and Shihua Wang; "Vector Quantization Techniques in Speech Coding"; Chapter 2 of Advances in Speech Signal Processing; edited by Sadoki Furui and Mohan Sondhi; Marcel Dekker, Inc.; New York, New York; 1992; pp.49-84
Kroon, Peter and Bishnu S. Atal; "Predictive Coding of Speech Using Analysis-by-Synthesis Techniques"; Chapter 5 of Advances in Speech Signal Processing; edited by Sadaoki Furui and M. Mohan Sondhi; Marcel Dekker, Inc.; New York, New York; 1992; pp. 141-164
Honda, Masaaki and Yoshinao Shiraki; "Very Low-Bit-Rate Speech Coding"; Chapter 7 of Advances in Speech Signal Processing; edited by Sadaoki Furui and M. Mohan Sondhi; Marcel Dekker, Inc.; New York, New York; 1992; pp. 209-230
Schroeter, Juergen and M. Mohan Sondhi; "Speech Coding Based on Physiological Models of Speech Production"; Chapter 8 of Advances in Speech Signal Processing; edited by Sadaoki Furui and M. Mohan Sondhi; Marcel Dekker, Inc.; New York, New York; 1992; pp. 231-26
August 6, 2017. Estimated Expiration Date is calculated based on simple USPTO term provisions. It does not account for terminal disclaimers, term adjustments, failure to pay maintenance fees, or other factors which might affect the term of a patent.
