Landmark-Guided Cross-Speaker Lip Reading with Mutual Information Regularization.
Linzhi WuXingyu ZhangYakun ZhangChangyan ZhengTiejun LiuLiang XieYe YanErwei YinPublished in: LREC/COLING (2024)
Keyphrases
- mutual information
- lip reading
- speaker identification
- image registration
- visual speech
- head tracking
- expression recognition
- speech recognition
- gaussian mixture model
- similarity measure
- feature selection
- image processing
- speech signal
- hidden markov models
- machine learning
- feature extraction
- automatic speech recognition
- noisy environments
- speaker verification
- computer vision
- broadcast news
- pattern recognition
- human computer interaction
- maximum likelihood
- image analysis