Multiple time resolution analysis of speech signal using MCE training with application to speech recognition.
Spiros DimopoulosAlexandros PotamianosEric Fosler-LussierChin-Hui LeePublished in: ICASSP (2009)
Keyphrases
- speech recognition
- speech signal
- automatic speech recognition
- hidden markov models
- spectral analysis
- speaker identification
- noisy environments
- language model
- cepstral coefficients
- isolated word
- linear prediction
- speech synthesis
- speech recognizer
- background noise
- vocal tract
- speaker independent
- noisy speech
- speaker recognition
- acoustic features
- discriminative training
- speech enhancement
- speech recognition systems
- speech quality
- support vector machine
- feature selection