Uncorrelated Multilinear Principal Component Analysis (UMPCA)

The codes implement the Uncorrelated Multilinear Principal Component Analysis (UMPCA) algorithm.

Matlab source codes for Uncorrelated Multilinear Principal Component Analysis (UMPCA)

%[Algorithm]%

The matlab codes provided here implement the UMPCA algorithm presented in the
paper "UMPCA_TNN09.pdf" included in this package:

Haiping Lu, K.N. Plataniotis, and A.N. Venetsanopoulos,
"Uncorrelated Multilinear Principal Component Analysis for Unsupervised
Multilinear Subspace Learning",
IEEE Transactions on Neural Networks,
Vol. 20, No. 11, Page: 1820-1836, Nov. 2009.

"maxeig.m" (by Todd K. Moon) is a function used by "UMPCA.m' to get the leading
eigenvector.
---------------------------

%[Usages]%

Please refer to the comments in the codes for example usage on 2D data
"FERETC70A15S8_80x80" in the directory "FERETC70A15S8", which is used in the
paper above. Various partitions used in the paper are included in the directory
"FERETC70A15S8" for L=1 to 7.

Directory "USFGait17_32x22x10" contains the gait data used in the paper above.
---------------------------

%[Toolbox needed]%:

This code needs the tensor toolbox available at
http://csmr.ca.sandia.gov/~tgkolda/TensorToolbox/
This package includes tensor toolbox version 2.1 for convenience.
---------------------------

%[Restriction]%

In all documents and papers reporting research work that uses the matlab codes
provided here, the respective author(s) must reference the following paper:

[1] Haiping Lu, K.N. Plataniotis, and A.N. Venetsanopoulos,
"Uncorrelated Multilinear Principal Component Analysis for Unsupervised
Multilinear Subspace Learning",
IEEE Transactions on Neural Networks,
Vol. 20, No. 11, pp. 1820-1836, Nov. 2009.
---------------------------

%[Additional Resources]%

The BibTeX file "MPCApublications.bib" contains the BibTex for UMPCA and
related works. The included survey paper "SurveyMSL_PR2011.pdf" discusses the
relations between UMPCA and related works.