Iterative multi-channel coherence analysis with applications

Neural Netw. 2008 Mar-Apr;21(2-3):493-501. doi: 10.1016/j.neunet.2007.12.025. Epub 2007 Dec 28.

Abstract

An iterative learning algorithm for performing Multi-Channel Coherence Analysis (MCCA) is developed in this paper. MCCA is an extension of the well-known Canonical Correlation Analysis (CCA) that allows for more than two data channels to be analyzed. This paper discusses a standard method for performing MCCA and compares it to a newly developed data-driven and iterative approach. The proposed algorithm is then tested on two examples and its performance is evaluated in terms of estimation errors with respect to the values obtained using the standard MCCA algorithm. The first example uses a synthesized data set while the second example uses a real data set based on multi-spectral satellite imagery of the Earth's surface.

Publication types

  • Research Support, U.S. Gov't, Non-P.H.S.

MeSH terms

  • Algorithms*
  • Humans
  • Image Processing, Computer-Assisted
  • Least-Squares Analysis
  • Multivariate Analysis*
  • Neural Networks, Computer*
  • Satellite Communications
  • Signal Processing, Computer-Assisted
  • Subtraction Technique