A Smoothed Matrix Multivariate Elliptical Distribution-Based Projection Method for Feature Extraction

Algorithm 1

Iterative Algorithm for solving α.

	Input: a dataset of n training image matrices{Xi^c}∈R^m×n, m = p×q, and an image matrix Y ∈ R^p×q
	Initialize t = 1, R = I ∈ R^n×n, T = e ∈ R^n, λ > 0, μ > 0 and ε = 1×e^−3
	while not converged do
(1)	Update αi(t+1) as follows:
(2)	Update Rt+1 and Tt+1 as follows:
(3)	t = t+1.
(4)	If ||αt+1−αt ||∞ ≤ ε, break.
	end while
	Output: the adjacent weight matrix α