Deep Learning-based Quantitative Steganalysis to Detect Motion Vector Embedding of HEVC Videos.
Xiongbo HuangYongjian HuYufei WangBeibei LiuShuowei LiuPublished in: DSC (2020)
Keyphrases
- motion vectors
- deep learning
- video sequences
- video coding standard
- video data
- motion compensation
- motion estimation
- video coding
- video compression
- coding efficiency
- reference frame
- discrete cosine transform
- block matching
- motion estimation algorithm
- unsupervised learning
- video frames
- macroblock
- data hiding
- bit rate
- motion field
- video codec
- machine learning
- compressed domain
- motion compensated
- video content
- computational complexity
- block size
- intra prediction
- image sequences
- key frames
- inter frame
- spatial domain
- low complexity
- jpeg images
- error concealment
- transform domain
- mental models
- multiscale
- coding method
- weakly supervised
- dct coefficients
- image compression
- stereoscopic video