4th International Conference on Spoken Language Processing

Philadelphia, PA, USA
October 3-6, 1996

Japanese Large-vocabulary Continuous-speech Recognition using a Business-newspaper Corpus

Tatsuo Matsuoka (1), Katsutoshi Ohtsuki (2), Takeshi Mori (3), Sadaoki Furui (1,3), Katsuhiko Shirai (2)

(1) NTT Human Interface Laboratories; (2) Waseda University; (3) Tokyo Institute of Technology, Tokyo, Japan

We studied Japanese large-vocabulary continuous-speech recognition (LV CSR) for a Japanese business newspaper. To enable word N-grams to be used, sentences were first segmented into words (morphemes) using a morphological analyzer. Newspaper articles for about five years were used to train N-gram language models. To evaluate our recognition system, we recorded speech data for sentences from another set of articles. Using the speech corpus, LV CSR experiments were conducted. For 7k vocabulary, the word error rate was 82.8% when no grammar and context-independent acoustic models were used. This improved to 20.0% when both bigram language models and context-dependent acoustic models were used.

Full Paper

Bibliographic reference.  Matsuoka, Tatsuo / Ohtsuki, Katsutoshi / Mori, Takeshi / Furui, Sadaoki / Shirai, Katsuhiko (1996): "Japanese large-vocabulary continuous-speech recognition using a business-newspaper corpus", In ICSLP-1996, 22-25.