Adaptive objective function of ICA by Gaussian approximation in second-order polynomial feature space.
Yoshitatsu MatsudaKazunori YamaguchiPublished in: IJCNN (2016)
Keyphrases
- feature space
- objective function
- polynomial kernels
- feature extraction
- independent component analysis
- principal component analysis
- randomized approximation
- higher order
- linear computational complexity
- independent components
- kernel function
- mean shift
- linear programming
- feature vectors
- lagrange interpolation
- weighted sum
- gaussian kernels
- gaussian convolution
- approximation algorithms
- differential operators
- image representation
- statistically independent
- high order
- gaussian mixture model
- feature selection
- optimization problems
- classification accuracy
- lower bound
- optimal solution
- input space
- signal processing
- feature set
- dimensionality reduction
- cost function
- image retrieval
- low degree
- multiscale
- taylor series expansion
- face recognition