Twin-HMM-based audio-visual speech enhancement.
Ahmed Hussen AbdelazizSteffen ZeilerDorothea KolossaPublished in: ICASSP (2013)
Keyphrases
- audio visual
- speech enhancement
- hidden markov models
- noisy environments
- noise reduction
- speech signal
- multi modal
- signal to noise ratio
- single channel
- multi stream
- linear prediction
- speech recognition
- visual information
- visual data
- vocal tract
- multimedia
- wiener filter
- multi channel
- sound source
- audio features
- automatic speech recognition
- background noise
- image restoration
- visual features
- smoothing algorithm
- three dimensional
- information retrieval
- speech synthesis
- image coding
- multiresolution