A Hybrid Approach to Combining Conventional and Deep Learning Techniques for Single-Channel Speech Enhancement and Recognition.
Yanhui TuIvan TashevShuayb ZararChin-Hui LeePublished in: ICASSP (2018)
Keyphrases
- single channel
- speech enhancement
- deep learning
- multi channel
- noisy environments
- unsupervised learning
- object recognition
- noise reduction
- independent component analysis
- machine learning
- frequency domain
- wiener filter
- prior information
- wavelet decomposition
- smoothing algorithm
- pattern recognition
- speech signal
- signal to noise ratio
- vocal tract
- linear prediction
- brain activity
- speech recognition
- high frequency
- high resolution
- visual attention
- pairwise