A subspace approach for enhancing speech corrupted by colored noise.
Yi HuPhilipos C. LoizouPublished in: ICASSP (2002)
Keyphrases
- noise free
- additive gaussian noise
- noisy environments
- noise level
- noise reduction
- speech recognition
- feature space
- signal subspace
- signal to noise ratio
- random noise
- input data
- background noise
- missing data
- broadcast news
- salt pepper
- additive noise
- speech enhancement
- poisson noise
- subspace clustering
- low dimensional
- automatic speech recognition
- speech signal
- recognition engine
- noise model
- high dimensional data
- principal components
- noisy data
- data sets
- removal of impulse noise
- spoken language
- image noise
- arbitrary shape
- gaussian noise
- median filter
- audio visual
- dimensionality reduction
- principal component analysis