Speech enhancement for bone-conducted speech based on low-order cepstrum restoration.
Daiki WatanabeYosuke SugiuraTetsuya ShimamuraHisanori MakinaePublished in: ISPACS (2017)
Keyphrases
- speech enhancement
- low order
- wiener filter
- noise reduction
- high order
- noisy environments
- speech signal
- higher order
- signal to noise ratio
- single channel
- vocal tract
- linear combination
- blurred images
- point spread function
- image restoration
- markov models
- linear prediction
- noisy speech
- background noise
- speech recognition
- numerical solution
- smoothing algorithm
- synthetic and real images
- multi channel
- image processing
- basis functions
- additive noise
- spherical harmonics
- speech synthesis
- markov random field
- denoising
- prior information
- non stationary
- natural images
- small number