ISCA Archive ICSLP 1994
ISCA Archive ICSLP 1994

KT-stock: a speaker-independent large-vocabulary speech recognition system over the telephone

Myoung-Wan Koo, Sang-Kyu Park, Kyung-Tae Kong, Sam-joo Doh

In this paper, we present a speaker-independent, large-vocabulary speech recognition system over the telephone (KT-STOCK) . If a user say isolated Korean company name over the telephone according to the voice instruction, the system announces real time stock quotes. The system can recognize 892 words consisting of 9 control words and 883 company names listed on the Korea Stock Exchange. The system is a HMM(hidden Markov model)-based isolated speech recognizer which uses phoneme-like unit as a basic unit. Currently, we have achieved recognition results of 92.0% over the telephone network.