End-to-End Speaker Identification in Noisy and Reverberant Environments Using Raw Waveform Convolutional Neural Networks.
Daniele SalvatiCarlo DrioliGian Luca ForestiPublished in: INTERSPEECH (2019)
Keyphrases
- end to end
- speaker identification
- speech signal
- convolutional neural networks
- noisy environments
- speech recognition
- gaussian mixture model
- non stationary
- feature extraction
- automatic speech recognition
- multipath
- congestion control
- hidden markov models
- noise reduction
- neural network
- image classification
- face recognition
- high level
- multi modal
- noisy images
- broadcast news
- multimedia
- real world