article · KITopen · 2002

LVCSR-based language identification

Tanja Schultz, Ivica Rogina, Alex Waibel

Cited by 47Open access

Abstract

Automatic language identification is an important problem in building multilingual speech recognition and understanding systems. Building a language identification module for four languages we studied the influence of applying different levels of knowledge sources on a large vocabulary continuous speech recognition (LVCSR) approach, i.e. phonetic, phonotactic, lexical, and syntactic-semantic knowledge. The resulting language identification (LID) module can identify spontaneous speech input and can be used as a front end for the multilingual speech-to-speech translation system JANUS-II. A comparison of five LID systems showed that the incorporation of lexical and linguistic knowledge reduces the language identification error for the 2-language tests up to 50%. Based on these results we build a LID module for German, English, Spanish, and Japanese which yields 84% identification rate on the spontaneous scheduling task (SST).

Cite this paper

APA 7

Schultz, T., Rogina, I., & Waibel, A. (2002). LVCSR-based language identification. KITopen, 2, 781–784. https://doi.org/10.1109/icassp.1996.543237

MLA 9

Schultz, Tanja, et al. “LVCSR-based language identification.” KITopen, vol. 2, 2002, pp. 781–84, https://doi.org/10.1109/icassp.1996.543237.

Chicago

Schultz, Tanja, Ivica Rogina, and Alex Waibel. 2002. “LVCSR-based language identification.” KITopen 2: 781–84. https://doi.org/10.1109/icassp.1996.543237.

Harvard

Schultz, T., Rogina, I. and Waibel, A. (2002) ‘LVCSR-based language identification’, KITopen, 2, pp. 781–784. Available at: https://doi.org/10.1109/icassp.1996.543237.

BibTeX

@article{schultz2002lvcsr,
  title = {{LVCSR-based language identification}},
  author = {Schultz, Tanja and Rogina, Ivica and Waibel, Alex},
  journal = {KITopen},
  volume = {2},
  pages = {781--784},
  year = {2002},
  doi = {10.1109/icassp.1996.543237}
}

Made from the index record. Check names, capitalisation and page numbers against the paper. Need another style? AI for Research has over 10,000.