Title :
Methodology of speech recognition using island analysis
Author_Institution :
Electron. Eng. Div. II, Toyota Motor Corp., Shizuoka, Japan
Abstract :
Recently HMM (hidden Markov model) has been used as the way of the standard voice recognition but it is required to target more high recognition ratio by the expansion of the application of speech recognition. In this vocal study the shape analysis of the vocal wave inside the vocal period has been done toward its modified signals whether its voice is for a word or a simple sentence. And then the evaluation of this experiment by changing the category of the vocal parameters on shape has been done.
Keywords :
hidden Markov models; speech recognition; HMM; hidden Markov model; island analysis; recognition ratio; shape analysis; speech recognition methodology; vocal wave; voice recognition; Gravity; Hidden Markov models; Image processing; Image recognition; Shape; Signal analysis; Speech analysis; Speech recognition; Target recognition; Vocabulary;
Conference_Titel :
SICE 2002. Proceedings of the 41st SICE Annual Conference
Print_ISBN :
0-7803-7631-5
DOI :
10.1109/SICE.2002.1195772
