Additive noise compensation in the i-vector space for speaker recognition.
Waad Ben KhederDriss MatroufJean-François BonastreMoez AjiliPierre-Michel BousquetPublished in: ICASSP (2015)
Keyphrases
- vector space
- additive noise
- speaker recognition
- speech signal
- gaussian mixture model
- speaker identification
- maximum likelihood
- speech recognition
- noisy images
- spatial domain
- vector quantization
- speaker verification
- random variables
- probabilistic neural network
- similarity search
- distance measure
- feature vectors
- hidden markov models
- noisy environments
- non stationary
- automatic speech recognition
- edge detection
- multiscale
- low dimensional
- knn
- neural network
- em algorithm