DocumentCode :
1075168
Title :
Improved Phoneme-Based Myoelectric Speech Recognition
Author :
Zhou, Quan ; Jiang, Ning ; Englehart, Kevin ; Hudgins, Bernard
Author_Institution :
Inst. of Biomed. Eng., Univ. of New Brunswick, Fredericton, NB, Canada
Volume :
56
Issue :
8
fYear :
2009
Firstpage :
2016
Lastpage :
2023
Abstract :
This paper introduces an enhanced phoneme-based myoelectric signal (MES) speech recognition system. The system can recognize new words without retraining the phoneme classifier, which is considered to be the main advantage of phoneme-based speech recognition. It is shown that previous systems experience severe performance degradation when new words are added to a testing dataset. To maintain high accuracy with new words, several improvements are proposed. In the proposed MES speech recognition approach, the raw MES is processed by class-specific rotation matrices to spatially decorrelate the data prior to feature extraction in a preprocessing stage. Then, an uncorrelated linear discriminant analysis is used for dimensionality reduction. The resulting data are classified through a hidden Markov model classifier to obtain the phonemic log likelihoods of the phonemes, which are mapped to corresponding words using a word classifier. An average word classification accuracy of 98.533% is achieved over six subjects. The system offers dramatically improved accuracy when expanding a vocabulary, offering promise for robust large-vocabulary myoelectric speech recognition.
Keywords :
data reduction; electromyography; feature extraction; hidden Markov models; medical signal processing; pattern classification; speech recognition; class specific rotation matrices; data spatial decorrelattion; dimensionality reduction; feature extraction; hidden Markov model classifier; large vocabulary myoelectric speech recognition; phoneme based MES; phoneme classifier; phonemic log likelihoods; uncorrelated linear discriminant analysis; Acoustic noise; Biomedical engineering; Degradation; Facial muscles; Hidden Markov models; Linear discriminant analysis; Neck; Principal component analysis; Speech enhancement; Speech recognition; Vocabulary; Gaussian mixture model (GMM); hidden Markov model (HMM); linear discriminant analysis (LDA); myoelectric signal (MES); pattern classification; principal component analysis (PCA); speech recognition; uncorrelated LDA (ULDA); Electromyography; Facial Muscles; Humans; Muscle Contraction; Pattern Recognition, Automated; Speech; Speech Production Measurement; Speech Recognition Software;
fLanguage :
English
Journal_Title :
Biomedical Engineering, IEEE Transactions on
Publisher :
ieee
ISSN :
0018-9294
Type :
jour
DOI :
10.1109/TBME.2009.2024079
Filename :
5075581
Link To Document :
بازگشت