Autoencoder based multi-stream combination for noise robust speech recognition.
Sri Harish Reddy MallidiTetsuji OgawaKarel VeselýPhani S. NidadavoluHynek HermanskyPublished in: INTERSPEECH (2015)
Keyphrases
- speech recognition
- audio visual speech recognition
- noisy environments
- multi stream
- hidden markov models
- audio visual
- speaker identification
- language model
- speech signal
- automatic speech recognition
- background noise
- pattern recognition
- speech synthesis
- speech processing
- noisy speech
- speaker verification
- speech recognition systems
- speaker independent
- speech recognition technology
- neural network
- speech recognizer
- speech enhancement
- multi modal
- image classification
- noise reduction
- feature selection
- machine learning
- contextual information
- signal to noise ratio