Voice Disorder Classification Based on Multitaper Mel Frequency Cepstral Coefficients Features.
Ömer EskidereAhmet GürhanliPublished in: Comput. Math. Methods Medicine (2015)
Keyphrases
- mel frequency cepstral coefficients
- feature vectors
- feature set
- speaker recognition
- speaker identification
- spectral features
- extracting features
- speech signal
- gaussian mixture model
- principal component analysis
- speech recognition
- feature extraction
- classification accuracy
- acoustic features
- speaker diarization
- audio features
- linear predictive
- feature space
- extracted features
- pattern recognition
- feature selection
- vector quantization
- speaker verification
- probabilistic neural network
- training set
- cepstral coefficients
- low level
- image features
- decision trees
- non stationary
- audio signal
- hidden markov models
- unsupervised learning
- structural features