Feature extraction approach to blind source separation

Local independent component analysis is formulated as a task involving the extraction of local geometric structure in the joint distribution. Because the geometrical structure of statistical independence is not well captured by statistical descriptions such as moments and cumulants, we use feature d...

Full description

Saved in:
Bibliographic Details
Main Authors: Lin, J.K., Grier, D.G., Cowan, J.D.
Format: Conference Proceeding
Language:English
Subjects:
Blind source separation
Computer vision
Equations
Feature extraction
Histograms
Image analysis
Independent component analysis
Partitioning algorithms
Physics
Source separation
Online Access:Request full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Local independent component analysis is formulated as a task involving the extraction of local geometric structure in the joint distribution. Because the geometrical structure of statistical independence is not well captured by statistical descriptions such as moments and cumulants, we use feature detection tools from image analysis to locate the local independent component coordinate system. The resulting approach to source separation can be implemented in real time using conventional image analysis hardware. The generality of this approach is demonstrated by blind source separation of multi-modal sources, and the pseudo-separation of three sources from two mixtures.
ISSN:1089-3555
2379-2329
DOI:10.1109/NNSP.1997.622421