Codebook-based speech enhancement using Markov process and speech-presence probability.
Qi HeChangchun BaoFeng BaoPublished in: INTERSPEECH (2015)
Keyphrases
- speech enhancement
- markov process
- noisy environments
- noise reduction
- speech signal
- transition probabilities
- markov chain
- signal to noise ratio
- single channel
- vocal tract
- stochastic process
- linear prediction
- noisy speech
- speech recognition
- probability distribution
- stationary distribution
- wiener filter
- additive noise
- smoothing algorithm
- speech synthesis
- stochastic processes
- multi channel
- image representation
- random walk
- background noise
- markov models
- multiscale
- noisy images
- steady state
- state space
- hidden markov models
- feature vectors
- pattern recognition