Title :
Speech recognition using TES symbols
Author :
Joubert, L.J. ; du Toit, J.C. ; van Schalkwyk, J.J.D.
Author_Institution :
Dept. of Electron. & Comput. Eng., Pretoria Univ., South Africa
fDate :
8/30/1991 12:00:00 AM
Abstract :
This paper describes tests to determine whether time-encoded speech symbols can be used as a symbol feature set for speech recognition. For this purpose isolated word recognition (more specifically digit recognition) is investigated, using a vocabulary consisting of the 10 digits in Afrikaans. Three different speakers were used to compile the database. TES symbols were used to obtain classification results by means of both a Markov model and a neural network
Keywords :
Markov processes; encoding; neural nets; speech recognition; Afrikaans; Markov model; classification; database; digit recognition; isolated word recognition; neural network; speech recognition; time-encoded speech symbols; Databases; Electronic equipment testing; Encoding; Frequency; Hidden Markov models; Shape measurement; Signal processing; Speech processing; Speech recognition; Vocabulary;
Conference_Titel :
Communications and Signal Processing, 1991. COMSIG 1991 Proceedings., South African Symposium on
Conference_Location :
Pretoria
Print_ISBN :
0-7803-0040-8
DOI :
10.1109/COMSIG.1991.278234
