Embedding new observations via sparse-coding for non-linear manifold learning.
Bogdan RaducanuFadi DornaikaPublished in: Pattern Recognit. (2014)
Keyphrases
- manifold learning
- sparse coding
- sparse representation
- nonlinear dimensionality reduction
- manifold structure
- locality preserving projections
- dictionary learning
- high dimensional
- manifold embedding
- subspace learning
- geodesic distance
- dimensionality reduction
- natural images
- image classification
- low dimensional
- latent space
- feature space
- generative model
- unsupervised learning
- dimension reduction
- high dimensional data
- image patches
- linear combination
- face recognition
- semi supervised
- locally linear embedding
- image representation
- visual words
- test images
- neural network
- signal processing
- vector space
- pairwise
- pattern recognition
- support vector
- image segmentation
- data mining