Noise-Robust ASR for the third 'CHiME' Challenge Exploiting Time-Frequency Masking based Multi-Channel Speech Enhancement and Recurrent Neural Network.
Zaihu PangFengyun ZhuPublished in: CoRR (2015)
Keyphrases
- speech enhancement
- multi channel
- single channel
- recurrent neural networks
- noisy environments
- noise reduction
- neural network
- speech recognition
- automatic speech recognition
- frequency domain
- speech signal
- background noise
- signal processing
- artificial neural networks
- feed forward
- recurrent networks
- wiener filter
- independent component analysis
- image structure
- image quality
- denoising
- human visual system
- pattern recognition