Optimized feature set to assess acoustic perturbations in dysarthric speech.
Sunil NagarajaEduardo Castillo GuerraPublished in: INTERSPEECH (2009)
Keyphrases
- feature set
- mel frequency cepstral coefficients
- audio features
- acoustic features
- speech recognition systems
- speech recognition
- feature selection
- speech sounds
- speaker independent
- emotional speech
- feature space
- feature vectors
- random forest
- classification accuracy
- acoustic models
- feature extraction
- feature representation
- acoustic signal
- speech signal
- prosodic features
- speaker recognition
- linguistic features
- speech synthesis
- text to speech
- speech recognizer
- selected features
- feature subset
- genetic algorithm
- speaker identification
- broadcast news
- hidden markov models
- automatic speech recognition
- audio visual
- vowel phonemes
- machine learning
- vocal tract
- spectral features
- emotion recognition
- structural features
- semantic features
- feature selection algorithms