Cantonese syllable recognition using neural networks

This work describes a novel neural network based speech recognition system for isolated Cantonese syllables. Since Cantonese is a monosyllabic and ton al language, the recognition system is composed of two major components, na mely the tone recognizer and the base syllable recognizer. The tone recogni zer adopts the architecture of multilayer perceptron in which each output n euron represents a particular tone. The base syllable recognizer consists o f a large number of independently trained recurrent networks, each represen ting a designated Cantonese syllable. An integrated recognition algorithm i s developed to give the ultimate recognition results based on N-best output s of the two subrecognizers. To demonstrate the effectiveness of the propos ed methods, a speaker-dependent recognition system has been built with the vocabulary expanding progressively from 40 syllables to 200 syllables, In t he case of 200 syllables, a top-1 recognition accuracy of 81.8% has been at tained whilst the top-3 accuracy is 95.2%.