Abstract

This dissertation describes a number of algorithms developed to increase the robustness of automatic speech recognition systems with respect to changes in the environment. These algorithms attempt to improve the recognition accuracy of speech recognition systems when they are trained and tested in different acoustical environments, and when a desk-top microphone (rather than a close-talking microphone) is used for speech input. Without such processing, mismatches between training and testing conditions produce an unacceptable degradation in recognition accuracy. Two kinds of

Citations

6234	Maximum likelihood from incomplete data via the EM algorithm - Dempster, Laird, et al. - 1977
976	An algorithm for vector quantizer design - Linde, Buzo, et al. - 1980
666	An introduction to hidden Markov models - Rabiner, Juang - 1986
550	Comparison of parametric representations of monosyllabic word recognition in continuously spoken sentences - Davis, Mermelstein - 1980
515	Vector quantization - Gray - 1984
300	Suppression of Acoustic Noise in Speech using Spectral Subtraction - Boll - 1979
276	P.Noll, Digital Coding of Waveforms - Jayant - 1984
246	Speech enhancement using a minimum mean-square error short-time spectral amplitude estimator - Ephraïm, Malah - 1984
178	Cepstral analysis technique for automatic speaker verification - Furui - 1981
156	Some statistical issues in the comparison of speech recognition algorithms - Gillick, Cox - 1989
133	Effectiveness of linear prediction characteristics of the speech wave for automatic speaker identification and verification - Atal - 1974
132	Hidden Markov model decomposition of speech and noise - Varga, Moore - 1990
94	An overview of the SPHINX speech recognition system - Lee, Hon, et al. - 1990
93	Enhancement of speech corrupted by acoustic noise - Berouti, Schwartz, et al. - 1979
88	Modern Spectral Estimation - Kay - 1988
77	A joint synchrony/mean-rate model of auditory speech processing - Seneff - 1988
73	Computer-steered microphone arrays for sound transduction in large rooms - Flanagan, Johnson, et al. - 1985
71	Large-Vocabulary Speaker-Independent Continuous Speech Recognition: The SPHINX System - Lee - 1988
70	Parameter estimation of superimposed signals using the EM algorithm - Feder, Weinstein - 1988
59	All-Pole Modeling of Degraded Speech - Lim, Oppenheim - 1978
54	Subdivision of the audible frequency range into critical bands (frequenzgruppen - Zwicker - 1961
43	A comparison of several acoustic representations for speech recognition with degraded and undegraded speech - Hunt, Lefebvre - 1989
40	Application of an auditory model to speech recognition - Cohen - 1989
40	Continuous Speech Recognition Using Hidden Markov Modeles - Picone - 1990
39	Constrained iterative speech enhancement with application to speech recognition - Hansen, Clements - 1991
36	A microphone array with adaptive post-filtering for noise reduction in reverberant rooms - Zelinski - 1998
33	Discrete representation of signals - Oppenheim, Johnson - 1972
32	Multi-style training for robust isolated-word speech recognition - Lippmann, Martin - 1987
32	The SUMMIT speech recognition system: phonological modelling and lexical access - Zue, Glass, et al. - 1990
24	Speech recognition using noise-adaptive prototypes - Nadas, Nahamoo, et al. - 1989
23	A Family of Distortion measures based upon projection operation for robust speech recognition - Mansour, Juang - 1988
22	Speech enhancement. c - Lim, ed - 1983
22	Optimal estimators for spectral restoration of noisy speech - Porter, Boll - 1984
17	A mimimum mean square error approach for speech enhancement - Ephraim - 1990
17	Modelling Non-Verbal Sounds for Speech Recognition - Ward - 1989
16	A Waibel, “Noise reduction using connectionist models - Tamura - 1988
15	Tukey, 'The Quefrency Alanysis of Time Series for Echoes - Bokert, Healy, et al.
14	Compernolle D. Noise adaptation in a hidden Markov model speech recognition system. Comput Speech Lang - Van - 1989
14	Speaker dependent and independent speech recognition experiments with an auditory model - Hunt, Lefebvre - 1988
14	Normalizations and selection of speech segments for speaker recognition scoring - Li, Porter - 1988
14	Evaluation of LPC spectral matching measures for phonetic unit recognition - Shikano - 1986
13	Unsupervised speaker adaptation method based on hierarchical spectral clustering - Furui - 1989
13	Hidden Markov models for - Huang, Ariki, et al. - 1990
13	Noise compensation algorithms for use with hidden Markov model based speech recognition - Varga, Moore, et al. - 1988
11	Comparative study of several distortion measures for speech recognition - Nocerino, Soong, et al. - 1985
11	An analysis of a noise reduction neural network - Tamura - 1989
10	Maximum likelihood noise cancellation using the EM algorithm - Feder, Oppenheim, et al. - 1989
9	On vocabulary-independent speech modeling - Hon, Lee - 1990
8	Switching adaptive filters for enhancing noisy and reverberant speech from microphone array recordings - Compernolle - 1990
8	Stress Compensation and Noise Reduction Algorithms for Robust Speech Recognition - Hansen, Clements - 1989