A minimum converted trajectory error (MCTE) approach to high quality speech-to-lips conversion.
Xiaodan ZhuangLijuan WangFrank K. SoongMark Hasegawa-JohnsonPublished in: INTERSPEECH (2010)
Keyphrases
- high quality
- error rate
- speech recognition
- speaker verification
- higher quality
- visual speech
- speech signal
- error bounds
- trajectory data
- image quality
- low quality
- audio visual
- prediction error
- feature extraction
- error analysis
- dialogue system
- global minimum
- speaker recognition
- speech synthesis
- vocal tract
- endpoint detection