%IM_RESIZE Fixed mapping for resizing object images
%
% B = IM_RESIZE(A,SIZE,METHOD)
% B = A*IM_RESIZE([],SIZE,METHOD)
% B = A*IM_RESIZE(SIZE,METHOD)
%
% INPUT
% A Dataset or datafile
% SIZE Desired size
% METHOD Method, see IMRESIZE
%
% OUTPUT
% B Dataset or datafile
%
% DESCRIPTION
% The objects stored as images in the dataset or datafile A are resized
% using the IMRESIZE command. Default METHOD is 'nearest' (nearest
% neighbor interpolation).
%
% A special method is 'preserve', which exactly copies the existing data,
% cuts it, or extends it with zeros as far as appropriate. In SIZE the
% desired output size has to be stored. Note that for proper use in
% PRTools third size parameter of multi-band images, e.g. 3 for RGB
% images, has to be supplied.
%
% In case SIZE is a scalar the default METHOD is 'preserve', which
% implies that the first SIZE samples are taken, which is useful if
% A is a 1-D signal. Otherwise the deafult METHOD is 'nearest'.
%
% SEE ALSO (<a href="http://37steps.com/prtools">PRTools Guide</a>)
% MAPPINGS, DATASETS, DATAFILES, IM2OBJ, DATA2IM
% 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
%DXD 24-8-2007
% I rewrote a part of this function. Now there are default values
% given, a bug is removed, and the identation is correct again.
function b = im_resize(varargin)
argin = shiftargin(varargin,'vector');
argin = setdefaults(argin,[],[16,16],[],[]);
if mapping_task(argin,'definition')
b = define_mapping(argin,'fixed');
b = setname(b,'Image resize');
else
[a,imsize,method,par] = deal(argin{:});
if isempty(method)
if length(imsize) == 1
method = 'preserve';
else
method = 'nearest';
end
end
if isa(a,'prdataset') % allows datafiles too
isobjim(a);
b = filtim(a,mfilename,{imsize,method},imsize); % prepare execution image by image
else
b = double(a);
if strcmp(method,'preserve') | strcmp(method,'preserve_bottom') | ...
strcmp(method,'preserve_centre') | strcmp(method,'preserve_top')
% copy the image pixel by pixel into a larger or smaller image
sizeb = size(b); % original size
if length(imsize) == 1
imsize = [1 imsize];
end
if length(sizeb) == 3 & length(imsize) == 2
imsize(3) = sizeb(3);
end
if length(imsize) ~= length(sizeb)
error('Desired images size should have as many dimension as data')
end
sizec = min(imsize,sizeb); % size of part to be copied
subs = cell(1,length(sizec)); % store indices in cell array
[subs{:}] = ind2sub(sizec,[1:prod(sizec)]);
subsb = subs; subsc = subs; % cell arrays for original and result
switch method
case 'preserve_bottom'
delc = imsize-sizeb;
case 'preserve_centre'
delc = round((imsize-sizeb)/2);
case {'preserve_top','preserve'}
delc = zeros(1,length(imsize));
end
for j=1:length(delc) % indices for
if delc(j) > 0 % result
subsc{j} = subs{j} + delc(j);
elseif delc(j) < 0 % and original
subsb{j} = subs{j} - delc(j);
end
end
Lc = sub2ind(imsize,subsc{:}); % corresponding linear coordinates of result
Lb = sub2ind(sizeb,subsb{:}); % corresponding linear coordinates of original
c = zeros(imsize); % embed result in zeros
c(Lc) = b(Lb); % copy
b = c; % store result
else % resizing the image using Matlab's imresize
if length(imsize) > 1
b = imresize(b,imsize(1:2),method);
elseif imsize > 1 & round(imsize) == imsize
b = imresize(a,[imsize imsize],method);
else
[m,n] = size(a);
b = imresize(a,round(imsize*[m,n]),method);
end
end
end
end
return