Title :
A study of speech recognition based on fuzzy statistics
Author :
Sheng-li, Li ; Chao-huan, Hou
Author_Institution :
Inst. of Acoust., Acad. Sinica, Beijing, China
Abstract :
This paper describes a fuzzy statistics-based approach to speech recognition. We call it fuzzy statistics-based hidden Markov models (FSHMMs), the parameters of which are calculated by fuzzy statistics instead of by the traditional Baum-Welch algorithm. The benefit of using FSHMMs lies in two aspects: the output distribution density can match the speech data more accurately, and the models can be adapted for new data conveniently while the loss of representation for old data is small. The corresponding training procedure and decision rule of FSHMMs are also described. We do not prove that the training algorithm converges, but experimental evidence suggests that it does. In recognition tests conducted on an isolated complete vocabulary of Mandarin words (speaker-dependent system), a 90.3% correct word recognition was observed for a male speaker and 16% for a female speaker using the male speaker models. After being adapted by a few female data, the speaker-dependent system could raise the recognition accuracy by 127% for a female and only reduce the accuracy by 4.7% for the male speaker
Keywords :
fuzzy set theory; hidden Markov models; speech recognition; statistical analysis; Mandarin words; correct word recognition; curve fitting algorithm; decision rule; experiment; female data; female speaker; fuzzy set; fuzzy statistics; hidden Markov models; isolated complete vocabulary; male speaker; male speaker models; model adaptation; output distribution density; recognition accuracy; recognition tests; speaker dependent system; speech data; speech recognition; training algorithm; training procedure; Chaos; Curve fitting; Density functional theory; Fuzzy sets; Hidden Markov models; Histograms; Maximum likelihood estimation; Speech recognition; Statistical distributions; Statistics;
Conference_Titel :
Signal Processing, 1996., 3rd International Conference on
Conference_Location :
Beijing
Print_ISBN :
0-7803-2912-0
DOI :
10.1109/ICSIGP.1996.567373