% RIDGEFILTER - enhances fingerprint image via oriented filters
%
% Function to enhance fingerprint image via oriented filters
%
% Usage:
% newim = ridgefilter(im, orientim, freqim, kx, ky, showfilter)
%
% Arguments:
% im - Image to be processed.
% orientim - Ridge orientation image, obtained from RIDGEORIENT.
% freqim - Ridge frequency image, obtained from RIDGEFREQ.
% kx, ky - Scale factors specifying the filter sigma relative
% to the wavelength of the filter. This is done so
% that the shapes of the filters are invariant to the
% scale. kx controls the sigma in the x direction
% which is along the filter, and hence controls the
% bandwidth of the filter. ky controls the sigma
% across the filter and hence controls the
% orientational selectivity of the filter. A value of
% 0.5 for both kx and ky is a good starting point.
% showfilter - An optional flag 0/1. When set an image of the
% largest scale filter is displayed for inspection.
%
% Returns:
% newim - The enhanced image
%
% See also: RIDGEORIENT, RIDGEFREQ, RIDGESEGMENT
% Reference:
% Hong, L., Wan, Y., and Jain, A. K. Fingerprint image enhancement:
% Algorithm and performance evaluation. IEEE Transactions on Pattern
% Analysis and Machine Intelligence 20, 8 (1998), 777 789.
% Peter Kovesi
% www.peterkovesi.com
%
% January 2005
function newim = ridgefilter(im, orient, freq, kx, ky, showfilter)
if nargin == 5
showfilter = 0;
end
angleInc = 3; % Fixed angle increment between filter orientations in
% degrees. This should divide evenly into 180
im = double(im);
[rows, cols] = size(im);
newim = zeros(rows,cols);
[validr,validc] = find(freq > 0); % find where there is valid frequency data.
ind = sub2ind([rows,cols], validr, validc);
% Round the array of frequencies to the nearest 0.01 to reduce the
% number of distinct frequencies we have to deal with.
freq(ind) = round(freq(ind)*100)/100;
% Generate an array of the distinct frequencies present in the array
% freq
unfreq = unique(freq(ind));
% Generate a table, given the frequency value multiplied by 100 to obtain
% an integer index, returns the index within the unfreq array that it
% corresponds to
freqindex = ones(100,1);
for k = 1:length(unfreq)
freqindex(round(unfreq(k)*100)) = k;
end
% Generate filters corresponding to these distinct frequencies and
% orientations in 'angleInc' increments.
filter = cell(length(unfreq),180/angleInc);
sze = zeros(length(unfreq),1);
for k = 1:length(unfreq)
sigmax = 1/unfreq(k)*kx;
sigmay = 1/unfreq(k)*ky;
sze(k) = round(3*max(sigmax,sigmay));
[x,y] = meshgrid(-sze(k):sze(k));
reffilter = exp(-(x.^2/sigmax^2 + y.^2/sigmay^2)/2)...
.*cos(2*pi*unfreq(k)*x);
% Generate rotated versions of the filter. Note orientation
% image provides orientation *along* the ridges, hence +90
% degrees, and imrotate requires angles +ve anticlockwise, hence
% the minus sign.
for o = 1:180/angleInc
filter{k,o} = imrotate(reffilter,-(o*angleInc+90),'bilinear','crop');
end
end
if showfilter % Display largest scale filter for inspection
figure(7), imshow(filter{1,end},[]); title('filter');
end
% Find indices of matrix points greater than maxsze from the image
% boundary
maxsze = sze(1);
finalind = find(validr>maxsze & validrmaxsze & validc maxorientindex);
orientindex(i) = orientindex(i)-maxorientindex;
% Finally do the filtering
for k = 1:length(finalind)
r = validr(finalind(k));
c = validc(finalind(k));
% find filter corresponding to freq(r,c)
filterindex = freqindex(round(freq(r,c)*100));
s = sze(filterindex);
newim(r,c) = sum(sum(im(r-s:r+s, c-s:c+s).*filter{filterindex,orientindex(r,c)}));
end