Semi-supervised and active learning through Manifold Reciprocal kNN Graph for image retrieval.
Daniel Carlos Guimarães PedronetteYing WengAlexandro BaldassinChaohuan HouPublished in: Neurocomputing (2019)
Keyphrases
- knn
- semi supervised
- active learning
- graph construction
- neighborhood graph
- k nearest neighbor
- image retrieval
- graph embedding
- manifold learning
- semi supervised learning
- relevance feedback
- semi supervised classification
- label propagation
- nearest neighbor
- unlabeled data
- graph laplacian
- labeled data
- text categorization
- k nearest neighbour
- support vector machine svm
- hierarchical document clustering
- supervised learning
- labeled and unlabeled data
- similarity search
- distance function
- text classification
- knn algorithm
- machine learning
- nearest neighbor search
- manifold structure
- training set
- subspace learning
- neural network
- metric learning
- support vector machine
- pairwise
- feature space
- feature selection
- k nearest
- image annotation
- shows significant improvements
- knn classifier
- laplacian eigenmaps
- high dimensional
- pairwise constraints