MPE4G: Multimodal Pretrained Encoder for Co-Speech Gesture Generation.
Gwantae KimSeonghyeok NohInsung HamHanseok KoPublished in: CoRR (2023)
Keyphrases
- multimodal interfaces
- multi stream
- multimodal interaction
- human computer interaction
- audio visual
- learning mechanism
- human centered
- gesture recognition
- hand movements
- closely related
- speech recognition
- user interface
- belief networks
- rate distortion
- future trends
- probabilistic inference
- bit rate
- search algorithm
- multi modal
- motion estimation
- hidden markov models
- video codec
- bayesian networks
- weighted max sat
- hand gestures
- low complexity
- visual information
- image quality
- probabilistic model