High-Dimensional Spectral Feature Selection for 3D Object Recognition Based on Reeb Graphs.
Boyan BonevFrancisco EscolanoDaniela GiorgiSilvia BiasottiPublished in: SSPR/SPR (2010)
Keyphrases
- high dimensional
- feature selection
- high dimensionality
- dimensionality reduction
- graph classification
- reeb graph
- spectral decomposition
- feature space
- microarray data
- spectral methods
- low dimensional
- dimension reduction
- high dimension
- gene expression data
- small sample
- manifold learning
- laplacian matrix
- nearest neighbor
- variable selection
- feature selection algorithms
- similarity search
- object recognition
- text classification
- machine learning
- multi class
- spectral analysis
- mutual information
- directed graph
- support vector
- graph mining
- view independent
- irrelevant features
- d objects
- discriminative features
- unsupervised learning
- graph theory
- feature subset
- parameter space
- high dimensional data
- text categorization
- model selection
- information gain
- hyperspectral
- critical points
- graph representation
- graph matching
- support vector machine
- data sets
- kernel methods
- kernel function
- feature extraction