On the Choice of Perception Loss Function for Learned Video Compression.
Sadaf SalehkalaibarBuu PhanJun ChenWei YuAshish KhistiPublished in: CoRR (2023)
Keyphrases
- loss function
- video compression
- motion compensation
- video coding
- motion estimation
- video data
- pairwise
- video conferencing
- learning to rank
- support vector
- bit allocation
- motion compensated
- empirical risk
- discrete cosine transforms
- compression ratio
- distributed video coding
- stochastic gradient descent
- boosting framework
- motion vectors
- risk minimization
- image and video compression
- video sequences
- reproducing kernel hilbert space
- convex loss functions
- motion field
- boosting algorithms
- three dimensional
- solution path
- multimedia
- computer vision