A simple overcomplete ICA algorithm by non-orthogonal pair optimizations

Authors:
Yoshitatsu Matsuda;Kazunori Yamaguchi
Affiliations:
Department of Integrated Information Technology, Aoyama Gakuin University, Sagamihara-shi, Kanagawa, Japan;Department of General Systems Studies, The University of Tokyo
Venue:
IJCNN'09 Proceedings of the 2009 international joint conference on Neural Networks
Year:
2009

Citing 14
Cited 1

Independent component analysis, a new concept?

Signal Processing - Special issue on higher order statistics
High-order contrasts for independent component analysis

Neural Computation
Natural gradient learning for over- and under-complete bases in ICA

Neural Computation
Joint Approximate Diagonalization of Positive Definite Hermitian Matrices

SIAM Journal on Matrix Analysis and Applications
Estimating Overcomplete Independent Component Bases for Image Windows

Journal of Mathematical Imaging and Vision
Adaptive Blind Signal and Image Processing: Learning Algorithms and Applications

Adaptive Blind Signal and Image Processing: Learning Algorithms and Applications
A Variational Method for Learning Sparse and Overcomplete Representations

Neural Computation
Learning Overcomplete Representations

Neural Computation
Linear multilayer ICA generating hierarchical edge detectors

Neural Computation
A note on lewicki-sejnowski gradient for learning overcomplete representations

Neural Computation
Blind separation of independent sources for virtually any sourceprobability density function

IEEE Transactions on Signal Processing
Non-orthogonal joint diagonalization in the least-squares sensewith application in blind source separation

IEEE Transactions on Signal Processing
Blind separation of instantaneous mixtures of nonstationary sources

IEEE Transactions on Signal Processing
Fast and robust fixed-point algorithms for independent component analysis

IEEE Transactions on Neural Networks

Partial extraction of edge filters by cumulant-based ICA under highly overcomplete model

ICONIP'10 Proceedings of the 17th international conference on Neural information processing: models and applications - Volume Part II

Quantified Score

Hi-index	0.00

Visualization

Abstract

Overcomplete ICA is a method for solving blind source separation problems if the number of observed signals is less than that of source ones. In this paper, we propose an overcomplete ICA algorithm based on a simple contrast function which is defined as the sum of the covariances of the squares of signals over all the pairs. By applying non-orthogonal pair optimizations to the function, a simple ICA algorithm is derived. Theoretical analysis and numerical experiments suggest the validity of the proposed algorithm.