filterGauss

 n-dimensional Gaussian filter.

 Creates an image of a Gaussian with arbitrary covariance matrix. The
 dimensionality and size of the filter is determined by dims (eg dims=[10
 10] creates a 2D filter of size 10x10). If mu==[], it is calculated to be
 the center of the n-dim image.  C can be a full nxn covariance matrix, or
 an nx1 vector of variance.  In the latter case C is calculated as
 C=diag(C). If C=[]; then C=(dims/6).^2, ie it is transformed into a
 vector of variances such that along each dimension the variance is equal
 to (siz/6)^2.

 USAGE
  G = filterGauss( dims, [mu], [C], [show] )

 INPUTS
  dims    - n element vector of dimensions of final Gaussian
  mu      - [] n element vector specifying the mean
  C       - [] nxn cov matrix, nx1 set of vars, or variance
  show    - [0] figure to use for optional display

 OUTPUTS
  G       - image of the created Gaussian

 EXAMPLE
  g = filterGauss( 21, [], 4, 1); %1D
  sig=3; G = filterGauss( 4*[sig sig] + 1, [], [sig sig].^2, 2 ); %2D
  R = rotationMatrix( [1,1,0], pi/4 );
  C = R'*[10^2 0 0; 0 5^2 0; 0 0 16^2]*R;
  G3 = filterGauss( [51,51,51], [], C, 3 ); %3D

 See also NORMPDF2

 Piotr's Computer Vision Matlab Toolbox      Version 2.0
 Copyright 2014 Piotr Dollar.  [pdollar-at-gmail.com]
 Licensed under the Simplified BSD License [see external/bsd.txt]