Example-based speech enhancement with joint utilization of spatial, spectral & temporal cues of speech and noise.
Keisuke KinoshitaMarc DelcroixMehrez SoudenTomohiro NakataniPublished in: INTERSPEECH (2012)
Keyphrases
- speech enhancement
- spectral subtraction
- linear prediction
- noisy environments
- speech signal
- noise reduction
- signal to noise ratio
- noisy speech
- single channel
- speech recognition
- vocal tract
- wiener filter
- space time
- spectral analysis
- background noise
- speaker identification
- smoothing algorithm
- additive noise
- lossless compression
- hidden markov models
- image coding
- image sequences
- prediction error
- edge preserving
- automatic speech recognition
- speech synthesis
- wavelet decomposition
- subband
- non stationary
- prior knowledge
- pattern recognition