Homography-based low rank approximation of light fields for compression.
Xiaoran JiangMikael Le PenduReuben A. FarrugiaSheila S. HemamiChristine GuillemotPublished in: ICASSP (2017)
Keyphrases
- light field
- low rank approximation
- face recognition across pose
- multi view
- singular value decomposition
- low rank
- image formation
- subspace learning
- iterative algorithms
- kernel matrix
- super resolution
- stereo matching
- spectral clustering
- compression ratio
- multiple views
- image compression
- adjacency matrix
- fundamental matrix
- latent semantic indexing
- feature points
- reconstruction error
- nonnegative matrix factorization
- kernel methods
- three dimensional
- camera motion
- image restoration
- dimensionality reduction
- semi supervised
- high resolution
- high quality