Motion-Based Convolutional Neural Networks for Super-Resolution from Compressed Videos.
Han-Yuan ChangHsiang-Yu HsiehYih-Chuan LinPublished in: ICCE-Berlin (2023)
Keyphrases
- super resolution
- video super resolution
- convolutional neural networks
- real world video sequences
- low resolution
- motion estimation
- motion compensation
- high resolution
- single frame
- image reconstruction
- image restoration
- high resolution video
- low resolution images
- multi frame
- video sequences
- motion field
- high quality
- motion blur
- image super resolution
- resolution enhancement
- super resolved
- light field
- higher resolution
- motion model
- super resolution reconstruction
- high resolution images
- motion vectors
- depth map
- image patches
- image sequences
- optical flow
- space time
- video coding
- compressed domain
- visual perception
- spatial resolution
- single image super resolution
- dynamic scenes
- motion segmentation
- key frames
- motion analysis