DocumentCode :
2871319
Title :
Optimisation of neural models for speaker identification
Author :
Oglesby, J. ; Mason, J.S.
Author_Institution :
Dept. of Electr. & Electron. Eng., Univ. Coll. of Swansea, UK
fYear :
1990
fDate :
3-6 Apr 1990
Firstpage :
261
Abstract :
An approach to speaker recognition based on feedforward neural models is investigated. Each person known to the system has a personalized neural net that is trained to be active for only that person´s speech. By including speech from many people in the training data of each net this approach can directly model differences in people´s speech. The chosen architecture and amount of training performed is shown to strongly affect the recognition performance. Large models with two hidden layers are shown to be inferior to models with only a single hidden layer and fewer weights. Recognition performance is shown to be comparable to that of a vector quantization approach based on personalized codebooks. The neural approach outperforms the codebook system for small model sizes but does slightly less well for larger models. A multitransputer implementation used in the training phase is described. Near linear speedup is obtained by splitting the training data to given independent subtasks, and a dynamic allocation scheme is used to assign these tasks to processors
Keywords :
encoding; neural nets; speech recognition; transputers; dynamic allocation scheme; feedforward models; hidden layer; independent subtasks; linear speedup; model sizes; multitransputer implementation; neural models; personalized codebooks; personalized neural net; recognition performance; speaker identification; speaker recognition; training data; training phase; vector quantization approach; weights; Automatic speech recognition; Educational institutions; Feedforward systems; Neural networks; Pattern recognition; Security; Speaker recognition; Speech; Speech recognition; Training data; Vector quantization;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Acoustics, Speech, and Signal Processing, 1990. ICASSP-90., 1990 International Conference on
Conference_Location :
Albuquerque, NM
ISSN :
1520-6149
Type :
conf
DOI :
10.1109/ICASSP.1990.115617
Filename :
115617
Link To Document :
بازگشت