Audiovisual-to-articulatory speech inversion using Active Appearance Models for the face and Hidden Markov Models for the dynamics.
Athanassios KatsamanisGeorge PapandreouPetros MaragosPublished in: ICASSP (2008)
Keyphrases
- hidden markov models
- speech recognition
- speech signal
- multi stream
- visual speech
- expression recognition
- visual speech recognition
- speech synthesis
- automatic speech recognition
- vocal tract
- phoneme recognition
- speech processing
- audio visual
- speech recognizer
- markov models
- keyword spotting
- sequence classification
- conditional random fields
- facial motion
- speaker identification
- automatic speech recognition systems
- viterbi algorithm
- visual information
- sequential data
- facial expression recognition
- speaker independent
- human faces
- hidden states
- markov model
- facial expressions
- handwriting recognition
- speaker dependent
- text recognition
- noisy environments
- dynamical systems
- emotion recognition
- face images
- human computer interaction
- lip reading
- gesture recognition
- speaker adaptation
- pattern recognition
- facial features
- hidden state
- acoustic features
- bayesian networks