Advertisement

Impact of Age in ASR for the Elderly: Preliminary Experiments in European Portuguese

  • Thomas Pellegrini
  • Isabel Trancoso
  • Annika Hämäläinen
  • António Calado
  • Miguel Sales Dias
  • Daniela Braga
Part of the Communications in Computer and Information Science book series (CCIS, volume 328)

Abstract

Standard automatic speech recognition (ASR) systems use acoustic models typically trained with speech of young adult speakers. Ageing is known to alter speech production in ways that require ASR systems to be adapted, in particular at the level of acoustic modeling. This paper reports ASR experiments that illustrate the impact of speaker age on speech recognition performance. A large read speech corpus in European Portuguese allowed us to measure statistically significant performance differences among age groups ranging from 60- to 90-year-old speakers. An increase of 41% relative (11.9% absolute) in word error rate was observed between 60-65-year-old and 81-86-year-old speakers. This paper also reports experiments on retraining acoustic models (AMs), further illustrating the impact of ageing on ASR performance. Differentiated gains were observed depending on the age range of the adaptation data use to retrain the acoustic models.

Keywords

ASR Portuguese Elderly Speech 

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. 1.
    Wilpon, J., Jacobsen, C.: A study of speech recognition for children and the elderly. In: Proc. ICASSP, Atlanta, pp. 349–352 (1996)Google Scholar
  2. 2.
    Baba, A., Yoshizawa, S., Yamada, M., Lee, A., Shikano, K.: Acoustic models of the elderly for large-vocabulary continuous speech recognition. Electronics and Communications in Japan 87(7), 49–57 (2004)Google Scholar
  3. 3.
    Vipperla, R., Renals, S., Frankel, J.: Longitudinal study of ASR performance on ageing voices. In: Proc. Interspeech, Brisbane, pp. 2550–2553 (2008)Google Scholar
  4. 4.
    Baeckman, L., Small, B., Wahlin, A.: Aging and memory: cognitive and biological perspectives. In: Handbook of the Psychology of Aging, pp. 349–377 (2001)Google Scholar
  5. 5.
    Fozard, J., Gordon-Salant, S.: Changes in vision and hearing with aging. In: Handbook of the Psychology of Aging, pp. 241–266 (2001)Google Scholar
  6. 6.
    Anderson, S., Liberman, N., Bernstein, E., Foster, S., Cate, E., Levin, B., Hudson, R.: Recognition of elderly speech and voice-driven document retrieval. In: Proc. ICASSP, Phoenix, pp. 145–148 (1999)Google Scholar
  7. 7.
    Neto, J., Meinedo, H., Viveiros, M., Cassaca, R., Martins, C., Caseiro, D.: Broadcast news subtitling system in portuguese. In: Proc. ICASSP 2008, Las Vegas, USA (2008)Google Scholar
  8. 8.
    Meinedo, H.: Audio pre-processing and speech recognition for broadcast news. Ph.D. dissertation, IST, Lisbon, Portugal (2008)Google Scholar
  9. 9.
    Meinedo, H., Caseiro, D.A., Neto, J.P., Trancoso, I.: AUDIMUS.MEDIA: A Broadcast News Speech Recognition System for the European Portuguese Language. In: Mamede, N.J., Baptista, J., Trancoso, I., Nunes, M.d.G.V. (eds.) PROPOR 2003. LNCS, vol. 2721, pp. 9–17. Springer, Heidelberg (2003)CrossRefGoogle Scholar
  10. 10.
    Meinedo, H., Abad, A., Pellegrini, T., Neto, J., Trancoso, I.: The L2F Broadcast News Speech Recognition System. In: Proc. Fala, Vigo, pp. 93–96 (2010)Google Scholar
  11. 11.
    Abad, A., Neto, J.: Incorporating Acoustical Modelling of Phone Transitions in a Hybrid ANN/HMM Speech Recognizer. In: Proceedings of INTERSPEECH, Brisbane, pp. 2394–2397 (2008)Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2012

Authors and Affiliations

  • Thomas Pellegrini
    • 1
  • Isabel Trancoso
    • 1
    • 2
  • Annika Hämäläinen
    • 3
    • 4
  • António Calado
    • 3
  • Miguel Sales Dias
    • 3
    • 4
  • Daniela Braga
    • 3
    • 4
  1. 1.INESC-ID LisboaLisbonPortugal
  2. 2.Instituto Superior TécnicoLisbonPortugal
  3. 3.Microsoft Language Development CenterLisbonPortugal
  4. 4.ADETTI ISCTE, IULLisbonPortugal

Personalised recommendations