My lips are concealed: Audio-visual speech enhancement through obstructions.
Triantafyllos AfourasJoon Son ChungAndrew ZissermanPublished in: CoRR (2019)
Keyphrases
- audio visual
- speech enhancement
- noisy environments
- noise reduction
- multi modal
- signal to noise ratio
- single channel
- speech signal
- linear prediction
- visual information
- visual speech
- vocal tract
- visual data
- multi stream
- smoothing algorithm
- sound source
- wiener filter
- multimedia
- speaker verification
- multi channel
- information retrieval
- speech recognition
- audio features
- edge detection
- probabilistic model
- low level
- hidden markov models
- feature space
- video sequences
- high level
- computer vision