Title :
Parallel algorithms for automatic speech recognition
Author :
Arriola, Y. ; Carrasco, R.A.
Author_Institution :
Dept. of Electr. & Electron. Eng., Staffordshire Polytech., UK
Abstract :
In the research, neural nets, or to be more precise, multilayer perceptrons (MLP), are used for the nonlinear phonetic mapping of the acoustic features. These phonetic outputs are used, subsequently, for the final probabilistic computation of the hidden Markov models (HMM). The speech recognition system was implemented and tested on real speech, low-pass filtered with a cut-off frequency of 3.8 kHz, and sampled at 8 kHz
Keywords :
Markov processes; neural nets; parallel algorithms; speech recognition; 3.8 kHz; 8 kHz; acoustic features; automatic speech recognition; cut-off frequency; hidden Markov models; multilayer perceptrons; neural nets; nonlinear phonetic mapping; phonetic outputs; probabilistic computation; sampling;
Conference_Titel :
Techniques for Speech Processing, IEE Colloquium on
Conference_Location :
London
