Frequency Domain Multi-channel Acoustic Modeling for Distant Speech Recognition.
Minhua WuKen'ichi KumataniShiva SundaramNikko StrömBjörn HoffmeisterPublished in: ICASSP (2019)
Keyphrases
- speech recognition
- frequency domain
- multi channel
- single channel
- speech recognition systems
- spatial domain
- speech recognizers
- speaker independent
- hidden markov models
- acoustic models
- pattern recognition
- speech recognizer
- cross correlation
- power spectrum
- language model
- feature extraction
- noisy speech
- speech signal
- speaker identification
- automatic speech recognition
- speech synthesis
- denoising
- noisy environments
- subband
- neural network
- wavelet domain
- image processing
- information retrieval