MIMT: Masked Image Modeling Transformer for Video Compression.
Jinxi XiangKuan TianJun ZhangPublished in: ICLR (2023)
Keyphrases
- video compression
- block matching
- image data
- input image
- motion compensation
- discrete cosine transforms
- image retrieval
- video conferencing
- image content
- image segmentation
- motion compensated
- video coding
- image representation
- motion estimation
- compression rate
- vector field
- multiscale
- image authentication
- block matching algorithm
- video data
- computational complexity
- low bit rate
- image processing operations
- multimedia