A convolutional neural network approach for half-pel interpolation in video coding.
Ning YanDong LiuHouqiang LiFeng WuPublished in: ISCAS (2017)
Keyphrases
- video coding
- convolutional neural network
- motion estimation
- motion compensation
- motion vectors
- motion compensated
- rate distortion
- video compression
- face detection
- bit rate
- video quality
- video sequences
- video codec
- high definition
- block matching
- rate control
- motion compensated prediction
- motion estimation and compensation
- computational complexity
- motion field
- image and video compression
- macroblock
- compression efficiency
- computer vision
- neural network
- low complexity
- video coder
- video coding standard
- prediction error
- block based motion estimation
- image sequences
- motion model
- inter frame
- object based coding
- optical flow
- error resilience
- scalable video coding
- block size
- coding efficiency
- reference frame
- temporal correlation
- visual quality
- subband
- image quality
- video encoder
- object detection
- low bit rate
- temporal redundancy
- error concealment
- bitstream