In this work, the speaker normalisation problem is afforded by two different techniques. The first one is based non a linear transformation of acoustic data. The second is based on a neural network trained using the back propagation algorithm. Two metric spaces for the speakers data representation are used. These are shown to improve the normalization performance. A Dynamic Time Warping recognition system is used to test the two approaches and compare the recognition performances
Multilayer Neural Networks for Speaker Normalization
Giuliani, Diego;
1995-01-01
Abstract
In this work, the speaker normalisation problem is afforded by two different techniques. The first one is based non a linear transformation of acoustic data. The second is based on a neural network trained using the back propagation algorithm. Two metric spaces for the speakers data representation are used. These are shown to improve the normalization performance. A Dynamic Time Warping recognition system is used to test the two approaches and compare the recognition performancesFile in questo prodotto:
Non ci sono file associati a questo prodotto.
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
