Can you give me some tips about the algorithm I should use in a fingerprint matching function?

2 views (last 30 days)
Ok, so I have four vectors of two fingerprint images: for every image I have a vector that stores the position of the ridge endings and a vector that stores the position of the bifurcations.
For two days I am trying to figure out an algorithm to compare this two images correctly. Just to use the position is every minutiae is useless because the fingerprints can be the same, but the position can change and that would no nothing good. I tried to use the distance in every image between every minutiae, store it in a vector and compare the two position vectors to each other. That, sometimes, works but overall is no good.
My last attempt looks like this:
I have the four vectors I told you about at the top. I compute the size of every vector so I can run a "for" structure to get the distance. Lets say:
% C,D - ridge endings vectors for img1/img2
% M,N - bifur vectors for img1/img2
Cn = numel(C); % nr. elements / ridge endings // img 1
Dn = numel(D); % nr. elements / ridge endings // img 2
Mn = numel(M); % nr. elements / bifur endings // img 1
Nn = numel(N); % nr. elements / bifur endings // img 2
Then I do this to compute the distance between the minutiae of every image:
for i=1:Cn
for j=1:Mn
dif = C(i) - M(j);
vecimg1(end+1) = dif; % distance vector between C and M elements / img1
end
end
for i=1:Dn
for j=1:Nn
dif = D(i) - N(j);
vecimg2(end+1) = dif; distance vector between D and N elements / img2
end
end
I know everything looks ugly and it's bad. I'm new to MatLab and I have this huge paper about fingerprint matching and I need to do it kinda quick. I need some help, please!

Accepted Answer

Image Analyst
Image Analyst on 9 Jun 2015
Why are C and M 1-D vectors? Isn't your fingerprint image a 2D image so the minutiae would be 2D (x,y) coordinates? Anyway, if you have the Statistics toolbox, you can use pdist() to calculate distance of every point to every other point.

More Answers (0)

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!