Informed single-channel speech separation using HMM-GMM user-generated exemplar source.
Qi WangWai Lok WooSatnam Singh DlayPublished in: IEEE ACM Trans. Audio Speech Lang. Process. (2014)
Keyphrases
- user generated
- single channel
- speech signal
- sound source
- speech enhancement
- speaker identification
- speech recognition
- multi channel
- hidden markov models
- automatic speech recognition
- gaussian mixture model
- social media
- noisy environments
- vocal tract
- web content
- user generated content
- sentiment analysis
- independent component analysis
- frequency domain
- user interests
- audio features
- audio visual
- non stationary
- prior information
- pattern recognition
- language model
- wiener filter
- brain activity
- noise reduction
- additive noise
- web pages
- visual features
- wavelet transform
- multiresolution
- multiscale
- feature extraction