Graph-based transformation manifolds for invariant pattern recognition with kernel methods.
Alexei PozdnoukhovSamy BengioPublished in: ICPR (3) (2006)
Keyphrases
- kernel methods
- pattern recognition
- machine learning
- feature space
- kernel function
- learning problems
- support vector
- kernel matrix
- support vector machine
- image processing
- feature extraction
- low dimensional
- dimensionality reduction
- reproducing kernel hilbert space
- manifold learning
- kernel matrices
- kernel ridge regression
- kernel pca
- computer vision
- multiple kernel
- kernel fisher discriminant analysis
- support vector machine svm
- euclidean space
- kernel trick
- classification accuracy
- data mining
- neural network
- input space
- multiple kernel learning
- kernel learning
- linear combination
- graph kernels
- active learning