nach oben

Cognitive Computation

Erschienen in:

01.12.2013

Auditory-Inspired Morphological Processing of Speech Spectrograms: Applications in Automatic Speech Recognition and Speech Enhancement

verfasst von: Joyner Cadore, Francisco J. Valverde-Albacete, Ascensión Gallardo-Antolín, Carmen Peláez-Moreno

Erschienen in: Cognitive Computation | Ausgabe 4/2013

Einloggen

Aktivieren Sie unsere intelligente Suche, um passende Fachinhalte oder Patente zu finden.

search-config

KI-gestützte Suche

Aus

Abstract

New auditory-inspired speech processing methods are presented in this paper, combining spectral subtraction and two-dimensional non-linear filtering techniques originally conceived for image processing purposes. In particular, mathematical morphology operations, like erosion and dilation, are applied to noisy speech spectrograms using specifically designed structuring elements inspired in the masking properties of the human auditory system. This is effectively complemented with a pre-processing stage including the conventional spectral subtraction procedure and auditory filterbanks. These methods were tested in both speech enhancement and automatic speech recognition tasks. For the first, time-frequency anisotropic structuring elements over grey-scale spectrograms were found to provide a better perceptual quality than isotropic ones, revealing themselves as more appropriate—under a number of perceptual quality estimation measures and several signal-to-noise ratios on the Aurora database—for retaining the structure of speech while removing background noise. For the second, the combination of Spectral Subtraction and auditory-inspired Morphological Filtering was found to improve recognition rates in a noise-contaminated version of the Isolet database.

Vorheriger Artikel Characterizing Neurological Disease from Voice Quality Biomechanical Analysis

Nächster Artikel Detecting Speech Polarity with High-Order Statistics

Sie haben noch keine Lizenz? Dann Informieren Sie sich jetzt über unsere Produkte:

Springer Professional "Wirtschaft+Technik"

Online-Abonnement

Mit Springer Professional "Wirtschaft+Technik" erhalten Sie Zugriff auf:

über 102.000 Bücher
über 537 Zeitschriften

aus folgenden Fachgebieten:

Automobil + Motoren
Bauwesen + Immobilien
Business IT + Informatik
Elektrotechnik + Elektronik
Energie + Nachhaltigkeit
Finance + Banking
Management + Führung
Marketing + Vertrieb
Maschinenbau + Werkstoffe
Versicherung + Risiko

Jetzt Wissensvorsprung sichern!

Jetzt informieren

Springer Professional "Technik"

Online-Abonnement

Mit Springer Professional "Technik" erhalten Sie Zugriff auf:

über 67.000 Bücher
über 390 Zeitschriften

aus folgenden Fachgebieten:

Automobil + Motoren
Bauwesen + Immobilien
Business IT + Informatik
Elektrotechnik + Elektronik
Energie + Nachhaltigkeit
Maschinenbau + Werkstoffe

Jetzt Wissensvorsprung sichern!

Jetzt informieren

Springer Professional "Wirtschaft"

Online-Abonnement

Mit Springer Professional "Wirtschaft" erhalten Sie Zugriff auf:

über 67.000 Bücher
über 340 Zeitschriften

aus folgenden Fachgebieten:

Bauwesen + Immobilien
Business IT + Informatik
Finance + Banking
Management + Führung
Marketing + Vertrieb
Versicherung + Risiko

Jetzt Wissensvorsprung sichern!

Jetzt informieren

See [47] for a discussion on the acoustic cues employed by humans in each of the levels.

To pave the way for later notation we are going to introduce our own names for the logarithmic scales m, z and ERB rate, respectively.

The basic quantity over which perception of sound is measured is sound pressure level which is a normalised, logarithmic sound pressure—unless otherwise noted log refers in this paper to base-10 logarithms— $L_p = 20\log\frac{p}{p_0} (\hbox{dB SPL})$ where p is sound pressure and $p_0 = 20 \; {\mu}Pa$ is the reference sound pressure, the lowest audible pressure for human ears at mid-frequencies. A related quantity is sound (intensity) level, a normalised, logarithmic intensity level $L_I = 10\log\frac{I}{I_0} (\hbox{dB SL})$ where I ∝ p ² is the acoustic intensity, an energy related quantity. When using I ₀ = 10⁻¹² N/m ² for reference, both levels can be equated and we drop the subindex:

$$ L = 20\log\frac{p}{p_0} (\hbox{dB SPL}) = 10\log\frac{I}{I_0} (\hbox{dB SL}) $$

Both dB SPL and dB SL will be simply notated as dB.

At least not with the reduced number of filterbanks suggested by psychoacoustical experiments.

Baker J. The Dragon system—an overview. IEEE Trans Acoust Speech Signal Process. 1975;23(1):24–29CrossRef

Beerends J, Hekstra A, Rix A, Hollier M. Perceptual evaluation of speech quality (PESQ) the new ITU standard for end-to-end speech quality assessment. Part II: psychoacoustic model. J Audio Eng Soc. 2002;50(10):765–78

Berouti M, Schwartz R, Makhoul J Enhancement of speech corrupted by acoustic noise. IEEE Int Conf Acoust Speech Signal Process. 1979;4:208–211. IEEE.

Bourlard H, Morgan N. Hybrid HMM/ANN systems for speech recognition: overview and new research directions. Adapt Process Seq Data Struct. 1998;389–417.

Cole R, Muthusamy Y, Fanty M. The isolet spoken letter database. 2011. http://www.cslu.ogi.edu/corpora/isolet.

Davis S, Mermelstein P. Comparison of parametric representations for monosyllabic word recognition in continuously spoken sentences. IEEE Trans Acoust Speech Signal Process. 1980;28(4):357–66CrossRef

Dougherty ER, Lotufo RA. Hands-on morphological image processing. Tutorial texts in optical engineering, vol. TT59. SPIE press 2003.

Ephraim Y, Malah D. Speech enhancement using a minimum mean-square error short-time spectral amplitude estimator. IEEE Trans Acoust Speech Signal Process. 1984;32(6):1109–21CrossRef

Evans N, Mason J, Roach M, et al. Noise compensation using spectrogram morphological filtering. In: Proceedings of the 4th IASTED International Conference on Signal and Image Processing. 2002. pp. 157–61.

10.

Ezeiza A, López de Ipiña K, Hernández C, Barroso N. Enhancing the feature extraction process for automatic speech recognition with fractal dimensions. Cogn Comput. 2012. pp. 1–6.

11.

Fastl H, Zwicker E. Psycho-acoustics: facts and models, 3rd edn. New York: Springer; 2007.

12.

Faundez-Zanuy M, Hussain A, Mekyska J, Sesa-Nogueras E, Monte-Moreno E, Esposito A, Chetouani M, Garre-Olmo J, Abel A, Smekal Z, López de Ipiña K. Biometric applications related to human beings: there is life beyond security. Cogn Comput. 2012; 1–16.

13.

Florentine M, Fastl H, Buus S. Temporal integration in normal hearing, cochlear impairment, and impairment simulated by masking. J Acoust Soc Am. 1998; 84(1):195–203.PubMedCrossRef

14.

Gelbart D, Hemmert W, Holmberg M, Morgan N. Noisy ISOLET and ISOLET testbeds. database. 2011. http://www.icsi.berkeley.edu/Speech/papers/eurospeech05-onset/isolet/.

15.

Glasberg B, Moore B. Derivation of auditory filter shapes from notched-noise data. Hear Res. 1990;47(1–2):103–38PubMedCrossRef

16.

Gonzalez R, Woods R Digital image processing. Boston: Addison-Wesley; 1993.

17.

Greenberg S. The integration of phonetic knowledge in speech technology, Text, Speech and Language Technology vol. 25, chap. From here to utility. New York: Springer; 2005. pp. 107–132.

18.

Gunawan TS, Ambikairajah E, Epps J. Perceptual speech enhancement exploiting temporal masking properties of human auditory system. Speech Commun. 2010;52:381–93CrossRef

19.

Hansen J, Pellom B. An effective quality evaluation protocol for speech enhancement algorithms. In: International Conference on Spoken Language Processing. Sydney, Australia; 1998. pp. 2819–22.

20.

Heckmann M, Domont X, Joublin F, Goerick C A hierarchical framework for spectro-temporal feature extraction. Speech Commun. 2010; (53):736–52.

21.

Hirsch H, Pearce D. The AURORA experimental framework for the performance evaluation of speech recognition systems under noisy conditions. In: ASR2000-Automatic Speech Recognition: Challenges for the new Millenium ISCA Tutorial and Research Workshop (ITRW) 2000.

22.

Hu Y, Loizou P. Evaluation of objective quality measures for speech enhancement. IEEE Trans Audio Speech Lang Process. 2008;16(1):229–38.CrossRef

23.

Hu Y, Loizou P. Evaluation of objective measures for speech enhancement. In: Proceedings of the Interspeech. 2006; pp. 1447–50 .

24.

Hurmalainen A, Virtanen T Modelling spectro-temporal dynamics in factorisation-based noise-robust automatic speech recognition. In: International Conference on Acoustics, Speech and Signal Processing (ICASSP), IEEE 2012; pp. 4113–16.

25.

Irino T, Patterson R A time-domain, level-dependent auditory filter: The gammachirp. J Acoust Soc Am 1997;101(1):412–19.CrossRef

26.

Irino T, Patterson R. A dynamic compressive gammachirp auditory filterbank. IEEE Trans Audio Speech Lang Process. 2006;14(6):2222–32CrossRef

27.

Jelinek F, Bahl L, Mercer R. Design of a linguistic statistical decoder for the recognition of continuous speech. IEEE Trans Inf Theory. 1975;21(3):250–56CrossRef

28.

Jesteadt W, Bacon SP, Lehman JR. Forward masking as a function of frequency, masker level, and signal delay. J Acoust Soc Am. 1982;71(4):950–62PubMedCrossRef

29.

Klatt D. Prediction of perceived phonetic distance from critical-band spectra: a first step. In: IEEE International Conference on Acoustics, Speech, and Signal Processing, vol. 7, 1982. pp. 1278–81.

30.

Loizou P. Matlab software. 2011. http://www.utdallas.edu/loizou/speech/software.htm.

31.

Martínez C, Goddard J, Milone D, Rufiner H. Bioinspired sparse spectro-temporal representation of speech for robust classification. Comput Speech Lang. 2012;26:336–48.CrossRef

32.

Matheron G, Serra J. The birth of mathematical morphology. In: Proceedings of the 6th International Symposium on Mathematical Morphology. Sydney, Australia; 2002. pp. 1–16.

33.

Meddis R. Simulation of mechanical to neural transduction in the auditory receptor. J Acoust Soc Am. 1986;79(3):702–11PubMedCrossRef

34.

Meddis R. Simulation of auditory-neural transduction: further studies. J Acoust Soc Am. 1988;83(3):1056–63PubMedCrossRef

35.

Meyer B, Kollmeier B. Robustness of spectro-temporal features against intrinsic and extrinsic variations in automatic speech recognition. Speech Commun. 2010;53:753–67

36.

Moore B, Glasberg B. Suggested formulae for calculating auditory-filter bandwidths and excitation patterns. J Acoust Soc Am. 1983;74:750.PubMedCrossRef

37.

Moore B, Glasberg B. A revised model of loudness perception applied to cochlear hearing loss. Hear Res. 2004;188(1–2):70–88PubMedCrossRef

38.

Patterson R, Robinson K, Holdsworth J, McKeown D, Zhang C, Allerhand M. Complex sounds and auditory images. Aud Physiol Percept 1992;83:429–46CrossRef

39.

Peláez-Moreno C, García-Moral A, Valverde-Albacete F. Analyzing phonetic confusions using formal concept analysis. J Acoust Soc Am. 2010;128(3):1377–90PubMedCrossRef

40.

Quackenbush S, Barnwell T, Clements M. Objective measures of speech quality. Upper Saddle River: Prentice Hall Englewood Cliffs; 1988.

41.

Quatieri TF (2002) Discrete-time speech signal processing. Principles and Practice. Signal Processing. Upper Saddle River: Prentice Hall; 2002.

42.

Rabiner L, Juang BH. Fundamentals of speech recognition. Signal Processing. Upper Saddle River: Prentice Hall; 1993.

43.

Rix A, Hollier M, Hekstra A, Beerends J. Perceptual evaluation of speech quality (PESQ), the new ITU standard for end-to-end speech quality assessment. Part I: Time-delay compensation. J Acoust Soc Am. 2002;50(10):755–64

44.

Scalart P, Filho J. Speech enhancement based on a priori signal to noise estimation. In: IEEE International Conference on Acoustics, Speech, and Signal Processing, IEEE 1986. pp. 629–32.

45.

Serra J, Soille P (eds). Mathematical morphology and its application to image processing. Computational imaging and vision. Kluwer Academic 1994.

46.

Stevens SS, Volkmann J, Newman EB. A scale for the measurement of the psychological magnitude of pitch. J Acoust Soc Am. 1937;8:185–90.CrossRef

47.

Summerfield Q, Culling J. Auditory segregation of competing voices: absence of effectes of FM or AM coherence. Philos Trans R Soc Lond. 1992;336:357–66CrossRef

48.

ten Bosch L, Kirchhoff K. Editorial note: Bridging the gap between human and automatic speech recognition. Speech Commun. 2007;49(5):331–5CrossRef

49.

Weiss NA, Hasset MJ. Introductory statistics. Addison- Wesley, Reading; 1993. pp. 407–08.

50.

Yeh J, Chen C. Auditory front-ends for noise-robust automatic speech recognition. In: 7th International Symposium on Chinese Spoken Language Process (ISCSLP), IEEE 2010. pp. 205–08.

51.

Yin H, Hohmann V, Nadeu C. Acoustic features for speech recognition based on gammatone filterbank and instantaneous frequency. Speech Commun. 2010;53:707–15.

52.

Zwicker E, Feldtkeller R. The ear as a communication receiver. Woodbury: Acoustical Society of America; 1999.

53.

Zwicker E, Jaroszewski A. Inverse frequency dependence of simultaneous tone-on-tone masking patterns at low levels. J Acoust Soc Am. 1982;71(6):1508–12.PubMedCrossRef

54.

Zwicker E, Terhardt E. Analytical expressions for critical-band rate and critical bandwidth as a function of frequency. J Acoust Soc Am. 1980;68:1523CrossRef

Titel: Auditory-Inspired Morphological Processing of Speech Spectrograms: Applications in Automatic Speech Recognition and Speech Enhancement
verfasst von: Joyner Cadore
Francisco J. Valverde-Albacete
Ascensión Gallardo-Antolín
Carmen Peláez-Moreno
Publikationsdatum: 01.12.2013
Verlag: Springer US
Erschienen in: Cognitive Computation / Ausgabe 4/2013
Print ISSN: 1866-9956
Elektronische ISSN: 1866-9964
DOI: https://doi.org/10.1007/s12559-012-9196-6

Springer Professional

Abstract

Bitte loggen Sie sich ein, um Zugang zu Ihrer Lizenz zu erhalten.

Sie haben noch keine Lizenz? Dann Informieren Sie sich jetzt über unsere Produkte:

Springer Professional "Wirtschaft+Technik"

Springer Professional "Technik"

Springer Professional "Wirtschaft"

Weitere Artikel der Ausgabe 4/2013

Children’s Emotion Recognition from Spontaneous Speech Using a Reduced Set of Acoustic and Linguistic Features

Moral Judgment, Human Motivation, and Neural Networks

Enhancing the Feature Extraction Process for Automatic Speech Recognition with Fractal Dimensions

Low-variance Multitaper Mel-frequency Cepstral Coefficient Features for Speech and Speaker Recognition Systems

A Perceptual Visual Feature Extraction Method Achieved by Imitating V1 and V4 of the Human Visual System

Voice Quality Modification Using a Harmonics Plus Noise Model

Premium Partner