Smooth Gmm Based Multi-Talker Spectral Conversion for Spectrally Degraded Speech.
Chuping LiuQian-Jie FuShrikanth S. NarayananPublished in: ICASSP (5) (2006)
Keyphrases
- speaker recognition
- gaussian mixture model
- speaker identification
- hyperspectral images
- spectral features
- voice activity detection
- speech recognition
- linear predictive coding
- speech signal
- text to speech
- linear prediction
- hyperspectral imagery
- em algorithm
- mel frequency cepstral coefficients
- speaker verification
- restored image
- spoken language
- automatic speech recognition
- gaussian mixture
- broadcast news
- probabilistic neural network
- noisy environments
- speech synthesis
- image restoration
- mixture model
- hearing impaired
- image processing