Large-scale image classification with trace-norm regularization.
Zaïd HarchaouiMatthijs DouzeMattis PaulinMiroslav DudíkJérôme MalickPublished in: CVPR (2012)
Keyphrases
- norm regularization
- convex optimization
- canonical correlation analysis
- regularization methods
- total variation
- matrix completion
- sparse representation
- loss function
- group lasso
- multiple kernel learning
- max margin
- feature selection
- image restoration
- low rank
- multi class
- variable selection
- image denoising
- edge preserving
- denoising
- regularization term
- super resolution
- kernel learning
- image classification
- efficient optimization
- least squares
- image processing
- solution path
- data sets