Integrating multiple observations for model-based single-microphone speech separation with conditional random fields.
Yu Ting YeungTan LeeCheung-Chi LeungPublished in: ICASSP (2012)
Keyphrases
- conditional random fields
- integrating multiple
- hidden markov models
- graphical models
- automatic speech recognition
- probabilistic model
- sequence labeling
- information extraction
- maximum entropy
- random fields
- markov random field
- generative model
- web page prediction
- higher order
- named entity recognition
- markov models
- speech recognition
- markov networks
- crf model
- image labeling
- semi markov
- pairwise
- structured prediction
- speech signal
- non stationary
- structured learning
- speaker diarization