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