Unsupervised manifold learning by correlation graph and strongly connected components for image retrieval.
Daniel Carlos Guimarães PedronetteRicardo da Silva TorresPublished in: ICIP (2014)
Keyphrases
- manifold learning
- strongly connected components
- semi supervised
- image retrieval
- discriminant projection
- shortest path
- connected components
- dimensionality reduction
- nonlinear dimensionality reduction
- low dimensional
- argumentation semantics
- high dimensional
- unsupervised learning
- feature space
- feature extraction
- supervised learning
- high dimensional data
- pairwise
- variable ordering
- image representation
- sparse representation
- binary images
- dimension reduction
- max sat
- manifold structure
- active learning
- visual features
- computer vision
- principal component analysis
- closed contours
- training set