How to detect ridge endings & bifurcations of fingerprint image?

Can anyone explain to me how this coding works on finding the ridge endings and bifurcation? Thank you.

%Minutiae extraction
s=size(thinned_image);
N=3;%window size
n=(N-1)/2;
r=s(1)+2*n;
c=s(2)+2*n;
double temp(r,c);   
temp=zeros(r,c);bifurcation=zeros(r,c);ridge=zeros(r,c);
temp((n+1):(end-n),(n+1):(end-n))=thinned_image(:,:);
outImg=zeros(r,c,3);%For Display
outImg(:,:,1) = temp .* 255;
outImg(:,:,2) = temp .* 255;
outImg(:,:,3) = temp .* 255;
for x=(n+1+10):(s(1)+n-10)
    for y=(n+1+10):(s(2)+n-10)
        e=1;
        for k=x-n:x+n
            f=1;
            for l=y-n:y+n
                mat(e,f)=temp(k,l);
                f=f+1;
            end
            e=e+1;
        end;
         if(mat(2,2)==0)
            ridge(x,y)=sum(sum(~mat));
            bifurcation(x,y)=sum(sum(~mat));
         end
    end;
end;

% RIDGE END FINDING
[ridge_x ridge_y]=find(ridge==2);
len=length(ridge_x);
%For Display
for i=1:len
    outImg((ridge_x(i)-3):(ridge_x(i)+3),(ridge_y(i)-3),2:3)=0;
    outImg((ridge_x(i)-3):(ridge_x(i)+3),(ridge_y(i)+3),2:3)=0;
    outImg((ridge_x(i)-3),(ridge_y(i)-3):(ridge_y(i)+3),2:3)=0;
    outImg((ridge_x(i)+3),(ridge_y(i)-3):(ridge_y(i)+3),2:3)=0;

      outImg((ridge_x(i)-3):(ridge_x(i)+3),(ridge_y(i)-3),1)=255;
      outImg((ridge_x(i)-3):(ridge_x(i)+3),(ridge_y(i)+3),1)=255;
      outImg((ridge_x(i)-3),(ridge_y(i)-3):(ridge_y(i)+3),1)=255;
      outImg((ridge_x(i)+3),(ridge_y(i)-3):(ridge_y(i)+3),1)=255;
  end

%BIFURCATION FINDING
[bifurcation_x bifurcation_y]=find(bifurcation==4);
len=length(bifurcation_x);
%For Display
for i=1:len
    outImg((bifurcation_x(i)-3):(bifurcation_x(i)+3),(bifurcation_y(i)-3),1:2)=0;
    outImg((bifurcation_x(i)-3):(bifurcation_x(i)+3),(bifurcation_y(i)+3),1:2)=0;
    outImg((bifurcation_x(i)-3),(bifurcation_y(i)-3):(bifurcation_y(i)+3),1:2)=0;
    outImg((bifurcation_x(i)+3),(bifurcation_y(i)-3):(bifurcation_y(i)+3),1:2)=0;

      outImg((bifurcation_x(i)-3):(bifurcation_x(i)+3),(bifurcation_y(i)-3),3)=255;
      outImg((bifurcation_x(i)-3):(bifurcation_x(i)+3),(bifurcation_y(i)+3),3)=255;
      outImg((bifurcation_x(i)-3),(bifurcation_y(i)-3):(bifurcation_y(i)+3),3)=255;
      outImg((bifurcation_x(i)+3),(bifurcation_y(i)-3):(bifurcation_y(i)+3),3)=255;

[Image Analyst]

Read this format your code and this how to get an answer and fix your post (like formatting, including an image, etc.)

Not sure of your algorithm. Is it one you got from section 21.6 here in Vision Bib? Or one of your own creation?

[Alice W.T]

Hi, this is the algorithm from https://www.mathworks.com/matlabcentral/fileexchange/31926-fingerprint-minutiae-extraction?s_tid=prof_contriblnk I don't understand the part I mentioned in the post. Can you please explain to me?Thank you =]

Answer 1

[Image Analyst]

Why don't you just skeletonize with bwmorph(), then use the 'endpoints' and 'branchpoints' options of bwmorph to find endpoints and bifurcations?