Pseudo-coherence-based MVDR beamformer for speech enhancement with ad hoc microphone arrays.
Vincent Mohammad TavakoliJesper Rindom JensenMads Græsbøll ChristensenJacob BenestyPublished in: ICASSP (2015)
Keyphrases
- speech enhancement
- minimum variance
- linear prediction
- signal to noise ratio
- speech signal
- automatic speech recognition
- image coding
- noisy environments
- lossless compression
- prediction error
- noise reduction
- subband
- speech recognition
- autoregressive
- single channel
- sound source
- vocal tract
- spectral analysis
- visual information
- background noise
- speaker diarization
- edge preserving
- multi channel
- filter bank
- feature vectors