Frame-wise speech extraction with recursive expectation maximization for partially deformable microphone arrays.
Weixin MengJian LiYuhai GeXiaodong LiChengshi ZhengPublished in: Digit. Signal Process. (2024)
Keyphrases
- expectation maximization
- automatic speech recognition
- em algorithm
- speech recognition
- speaker diarization
- mixture model
- visual information
- generative model
- pairwise
- speech signal
- maximum likelihood
- gaussian mixture model
- probabilistic model
- audio visual
- information extraction
- maximum a posteriori
- frame rate
- automatic extraction
- text to speech
- recognition engine
- video frames
- image processing
- deformable models
- probability density function
- datalog programs
- broadcast news
- speaker recognition
- speech synthesis
- hidden markov models
- k means
- endpoint detection