Tight Convex Relaxations for Vector-Valued Labeling.
Bastian GoldlueckeEvgeny StrekalovskiyDaniel CremersPublished in: SIAM J. Imaging Sci. (2013)
Keyphrases
- vector valued
- convex relaxation
- convex optimization
- multi label
- globally optimal
- multistage
- scale space
- active learning
- wavelet packet
- image segmentation
- reproducing kernel hilbert space
- multiple kernel learning
- optimization methods
- unsupervised learning
- text categorization
- wavelet transform
- prior knowledge
- light field
- image denoising
- machine learning
- image restoration
- graph cuts