%IM_GAUSS Gaussian filter of images stored in a dataset/datafile (Matlab)
%
%	B = IM_GAUSS(A,SX,SY,SHAPE)
%	B = A*IM_GAUSS([],SX,SY,SHAPE)
%	B = A*IM_GAUSS(SX,SY,SHAPE)
%
% INPUT
%   A     Dataset with object images dataset (possibly multi-band)
%   SX    Desired horizontal standard deviation for filter, default SX = 1
%   SY    Desired vertical standard deviation for filter, default SY = SX
%   SHAPE Desired shape, 'same' (default) or 'full', see CONV2
%
% OUTPUT
%   B     Dataset/datafile with Gaussian filtered images
%
% DESCRIPTION
% All, possibly multi-band, 2D images in A are Gaussian filtered using the
% Matlab command CONV2. In case DIPImage is available, IM_GAUSSF may be
% used instead for faster processing.
%
% SEE ALSO (<a href="http://37steps.com/prtools">PRTools Guide</a>)
% DATASETS, DATAFILES, IM_GAUSSF, FILTIM, CONV2

% Copyright: R.P.W. Duin, r.p.w.duin@37steps.com
% Faculty EWI, Delft University of Technology
% P.O. Box 5031, 2600 GA Delft, The Netherlands

function b = im_gauss(varargin)

	argin = shiftargin(varargin,'scalar');
  argin = setdefaults(argin,[],1,[],'same');
  if mapping_task(argin,'definition')
    b = define_mapping(argin,'fixed');
    b = setname(b,'Gaussian filter');
  else
    [a,sx,sy,shape] = deal(argin{:});
    sy = setdefaults({sy},sx);
    if isa(a,'prdataset') % allows datafiles too
      if sy == 0 % 1-D processing
        b = filtm(a,mfilename,{sx,0});
      else
        b = filtim(a,mfilename,{sx,sy});
      end
    elseif isa(a,'double')  % here we have a single image
      if sx == 0
        fx = 1;
      else
        rx = round(3*sx);
        fx = exp((-[-rx:1:rx].^2)/(2*sx*sx)); fx = fx/sum(fx);
      end
      if sy == 0
        fy = 1;
      else
        ry = round(3*sy);
        fy = exp((-[-ry:1:ry].^2)/(2*sy*sy)); fy = fy/sum(fy);
      end
      n = size(a,3);
      b = zeros(size(a));
      for j=1:n
        b = conv2(fy,fx,a(:,:,j),shape);
        %b(:,:,j) = conv2(fy,fx,a(:,:,j),'same');
      end
    end
  end
return