论文信息 - Selection of pronunciation variants in spontaneous speech: comparing the performance of man and machine

Selection of pronunciation variants in spontaneous speech: comparing the performance of man and machine

Dans cet article, les performances d'un outil de transcription automatique sont evaluees. L'outil de transcription est un reconnaisseur de parole continue (CSR) fonctionnant en mode de reconnaissance forcee. Pour l'evaluation les performances du CSR ont ete comparees a celles de neuf auditeurs experts. La machine et l'humain ont effectue exactement la meme tâche: decider si un segment etait present ou non dans 467 cas. Il s'est avere que les performances du CSR etaient comparables a celle des experts.

Helmer Strik | Judith M. Kessens | Catia Cucchiarini | Mirjam Wester

[1] C. Cucchiarini,et al. Phonetic transcription: a methodological and empirical study , 1993 .

[2] H. V. D. Heuvel,et al. /r/ deletion in Standard Dutch , 1995 .

[3] Toni Rietveld,et al. Statistical Techniques for the Study of Language and Language Behaviour , 1993 .

[4] L D Shriberg,et al. A procedure for phonetic transcription by consensus. , 1984, Journal of speech and hearing research.

[5] G. Booij. The Phonology of Dutch , 1995 .

[6] Jørgen Rischel,et al. Formal linguistics and real speech , 1992, Speech Commun..

[7] Lou Boves,et al. A spoken dialog system for the Dutch public transport information service , 1997, Int. J. Speech Technol..

[8] Peter Ladefoged. Knowing enough to analyze spoken languages , 1992, ICSLP.

[9] Marc Swerts,et al. On the controlled elicitation of spontaneous speech , 1992, Speech Commun..

[10] Helmer Strik,et al. Improving the performance of a Dutch CSR by modeling pronunciation variation , 1998 .