Effect of pronunciations on OOV queries in spoken term detection


Can D., Cooper E., Sethy A., White C., Ramabhadran B., SARAÇLAR M.

2009 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2009, Taipei, Tayvan, 19 - 24 Nisan 2009, ss.3957-3960, (Tam Metin Bildiri)

  • Yayın Türü: Bildiri / Tam Metin Bildiri
  • Doi Numarası: 10.1109/icassp.2009.4960494
  • Basıldığı Şehir: Taipei
  • Basıldığı Ülke: Tayvan
  • Sayfa Sayıları: ss.3957-3960
  • Anahtar Kelimeler: Speech Recognition, Speech Indexing and Retrieval, Spoken Term Detection, Weighted Finite State Transducers
  • Boğaziçi Üniversitesi Adresli: Evet

Özet

The spoken term detection (STD) task aims to return relevant segments from a spoken archive that contain the query terms whether or not they are in the system vocabulary. This paper focuses on pronunciation modeling for Out-of-Vocabulary (OOV) terms which frequently occur in STD queries. The STD system described in this paper indexes word-level and sub-word level lattices or confusion networks produced by an LVCSR system using Weighted Finite State Transducers (WFST).We investigate the inclusion of n-best pronunciation variants for OOV terms (obtained from letter-to-sound rules) into the search and present the results obtained by indexing confusion networks as well as lattices. The following observations are worth mentioning: phone indexes generated from sub-words represent OOVs well and too many variants for the OOV terms degrade performance if pronunciations are not weighted. ©2009 IEEE.