VRED: A Position-Velocity Recurrent Encoder-Decoder for Human Motion Prediction.
Hongsong WangJiashi FengPublished in: CoRR (2019)
Keyphrases
- motion prediction
- wyner ziv
- distributed video coding
- motion estimation
- rate distortion
- low complexity
- video codec
- video coding
- motion vectors
- bi directional
- decoding process
- short term
- video compression
- kalman filter
- transform domain
- bit rate
- optical flow
- error resilience
- rate allocation
- image sequences
- video quality
- video sequences
- bit plane
- inter frame
- motion compensation
- coding scheme
- computer vision