What Is Optimized in Convex Relaxations for Multilabel Problems: Connecting Discrete and Continuously Inspired MAP Inference.
Christopher ZachChristian HaneMarc PollefeysPublished in: IEEE Trans. Pattern Anal. Mach. Intell. (2014)
Keyphrases
- convex relaxation
- multi label
- convex optimization
- multistage
- globally optimal
- image classification
- graph cuts
- text categorization
- learning tasks
- optimization methods
- multiple kernel learning
- image annotation
- light field
- machine learning
- evolutionary algorithm
- dynamic programming
- optimization problems
- bayesian networks