| | |
Summary: Norwegian Speech Recognition for Telephone
Applications
Harald Lj en Ingunn Amdal Finn Tore Johansen
Norwegian Telecom Research
N-2007 Kjeller, Norway
ABSTRACT
In this paper we present a Norwegian tele-
phone speech database, TABU.0. We discuss
the database design speci cation and some ex-
periences with recording and labelling of the
database. We also present some preliminary re-
sults with a word-based recogniser trained on a
subset of the database.
1. INTRODUCTION
Automatic speech recognition has now reached a
level of development where it is rapidly moving
from the laboratories into real-world applications,
many of them over telephone lines. The most im-
portant single factor for successful speech recogni-
tion seems to be the speech database used to train
|
