Combined Multi-Channel NMF-Based Robust Beamforming for Noisy Speech Recognition.
Masato MimuraYoshiaki BandoKazuki ShimadaShinsuke SakaiKazuyoshi YoshiiTatsuya KawaharaPublished in: INTERSPEECH (2017)
Keyphrases
- speech recognition
- multi channel
- noisy environments
- speech enhancement
- speech signal
- single channel
- hidden markov models
- automatic speech recognition
- speech recognizer
- speech synthesis
- speech processing
- language model
- negative matrix factorization
- speaker identification
- pattern recognition
- speech recognition technology
- noise reduction
- background noise
- speech recognition systems
- neural network
- speaker dependent
- speech retrieval
- multiscale
- speaker independent
- acoustic models
- word recognition
- frequency domain
- probabilistic model