Joint estimation of DOA and speech based on EM beamforming.
Lae-Hoon KimMark Hasegawa-JohnsonGerasimos PotamianosVit LibalPublished in: ICASSP (2010)
Keyphrases
- joint estimation
- direction of arrival
- speech enhancement
- doa estimation
- expectation maximization
- speech recognition
- em algorithm
- speech signal
- noisy environments
- confidence level
- text to speech
- unsupervised learning
- speech synthesis
- recognition engine
- single channel
- audio visual
- mixture model
- signal subspace
- probabilistic model
- maximum likelihood
- endpoint detection
- vocal tract
- dialogue system
- spoken language
- language acquisition
- emotion recognition
- image segmentation
- automatic speech recognition
- generative model
- bayesian networks
- noise reduction
- signal to noise ratio
- gaussian mixture model
- spontaneous speech
- linear array
- frequency domain
- computer simulation
- least squares
- pattern recognition
- audio stream