Single-Channel Speech-Music Separation for Robust ASR With Mixture Models.
Cemil DemirMurat SaraclarAli Taylan CemgilPublished in: IEEE Trans. Speech Audio Process. (2013)
Keyphrases
- mixture model
- single channel
- speech enhancement
- noisy environments
- speech recognition
- automatic speech recognition
- gaussian mixture model
- speech signal
- multi channel
- generative model
- language model
- probabilistic model
- em algorithm
- model selection
- expectation maximization
- unsupervised learning
- maximum likelihood
- audio features
- sound source
- probability density function
- noise reduction
- information retrieval
- machine learning
- frequency domain