Portal de Periódicos da CAPES

Covariance estimation with limited training samples

1999; Institute of Electrical and Electronics Engineers; Volume: 37; Issue: 4 Linguagem: Inglês

10.1109/36.774728

1558-0644

S. Tadjudin, D. A. Landgrebe,

Spectroscopy and Chemometric Analyses

This paper describes a covariance estimator formulated under an empirical Bayesian setting to mitigate the problem of limited training samples in the Gaussian maximum likelihood (ML) classification for remote sensing. The most suitable covariance mixture is selected by maximizing the average leave-one-out log likelihood. Experimental results using AVIRIS data are presented.