A Speaker-Independent Continuous Speech Recognition System Using Biomimetic Pattern Recognition
نویسنده
چکیده
-—— In speaker-independent speech recognition. the disadvantage of the most di行used technology (HMMs.or Hidden Markov models)is not only the need of m any m ore training sam ples,but also long train tim e requirem ent. This PaPer describes the use of Biom im etic pattern recognition(BPR)in recognizing some mandarin continuous speech in a speaker-independent m anner. A speech database was developed for the course of study.The vocabulary of the database consists of 15 Chinese dish’s nam es.the length of each nam e is 4 Chinese words.N eu. ral networks(NNs)based on Multi—weight neuron(MW N) m odel are used to train and recognize the speech sounds. The num ber 0f M W N was investigated to achieve the op. tim al perfo rm ance of the NN s.based B PR .T his system , which is based on BPR and can carry out real tim e recognition reaches a recognition rate of 98.14% for the flrst option and 99.81% for the flrst two options to the Per. sons from different provinces of China speaking com m on Chinese speech.Experim ents were also carried on to evalu. ate Continuous density hidden Markov models(CDHM M), Dynamic time warping(DTw )and BPR for speech recognition. The Experim ent results show that BPR outper. form s CD HM M and D TW especialy in the cases of sam . pies ofa finite size. K ey words-·——Biom im etic recognition, Hidden M arkov time warping(DTW ). pattern recognition,Speech models(HMMs),Dynamic
منابع مشابه
Speaker Adaptation in Continuous Speech Recognition Using MLLR-Based MAP Estimation
A variety of methods are used for speaker adaptation in speech recognition. In some techniques, such as MAP estimation, only the models with available training data are updated. Hence, large amounts of training data are required in order to have significant recognition improvements. In some others, such as MLLR, where several general transformations are applied to model clusters, the results ar...
متن کاملSpeaker Adaptation in Continuous Speech Recognition Using MLLR-Based MAP Estimation
A variety of methods are used for speaker adaptation in speech recognition. In some techniques, such as MAP estimation, only the models with available training data are updated. Hence, large amounts of training data are required in order to have significant recognition improvements. In some others, such as MLLR, where several general transformations are applied to model clusters, the results ar...
متن کاملSpeaker Independent Speech Recognition Using Hidden Markov Models for Persian Isolated Words
متن کامل
Speaker Independent Speech Recognition Using Hidden Markov Models for Persian Isolated Words
متن کامل
Biomimetic Pattern Recognition
In speaker-independent speech recognition, the disadvantage of the most diffused technology ( Hidden Markov Models) is not only the need of many more training samples, but also long train time requirement. This paper describes the use of Biomimetic Pattern Recognition (BPR) in recognizing some Mandarin Speech in a speaker-independent manner. The vocabulary of the system consists of 15 Chinese d...
متن کامل