Multi-channel linear prediction-based speech dereverberation with low-rank power spectrogram approximation.
Ante JukicNasser MohammadihaToon van WaterschootTimo GerkmannSimon DocloPublished in: ICASSP (2015)
Keyphrases
- multi channel
- low rank
- linear prediction
- speech signal
- single channel
- speech recognition
- convex optimization
- missing data
- matrix factorization
- linear combination
- matrix completion
- semi supervised
- singular value decomposition
- image coding
- high dimensional data
- lossless compression
- high order
- non stationary
- spectral analysis
- prediction error
- autoregressive
- hidden markov models
- neural network
- subband
- signal processing
- motion estimation