A Deep-Learning-Based Error-Correction Method for Atmospheric Motion Vectors.
Hang CaoHongze LengJun ZhaoYanlai ZhaoChengwu ZhaoBaoxu LiPublished in: Remote. Sens. (2024)
Keyphrases
- deep learning
- motion vectors
- prediction error
- motion estimation
- motion compensation
- video sequences
- video coding
- block matching
- bit rate
- compressed domain
- unsupervised learning
- reference frame
- macroblock
- motion field
- computational complexity
- machine learning
- unsupervised feature learning
- video data
- error propagation
- motion compensated
- rate distortion
- weakly supervised
- mental models
- block size
- frame interpolation
- motion compensated frame interpolation
- object recognition
- three dimensional
- computer vision
- higher order
- image sequences