Blind Determination of the Signal to Noise Ratio of Speech Signals Based on Estimation Combination of Multiple Features.
Russell OnduskoMatthew MarbachA. McClellanRavi Prakash RamachandranLinda M. HeadMark C. HugginsBrett Y. SmolenskiPublished in: APCCAS (2006)
Keyphrases
- signal to noise ratio
- multiple features
- speech enhancement
- speech signal
- spectral subtraction
- noise reduction
- combining multiple
- speech recognition
- noisy environments
- data fusion
- noisy speech
- wiener filter
- feature vectors
- bit error rate
- multiple instance learning
- feature set
- automatic speech recognition
- image retrieval
- linear prediction
- feature space
- metric learning
- gaussian noise
- low level features
- visual information
- edge detection
- incremental learning
- multi modal
- background noise
- probabilistic model
- vocal tract
- hidden markov models
- active learning
- multiscale