Bifurcation and ending in an edge image - matlab

Question

I'm trying to find Ending and Bifurcation points of a binary edge image produced from an ear image.
I'm writing the code based on the algorithm defined in [this paper] (http://elcvia.cvc.uab.es/article/view/108). I've implemented the part that introduced the formula to find endings and bifurcation. However, I don't get the correct result for different pictures, It's like this formula can't find some of the bifurcations and endings.
this is my code :
stats = regionprops(BW4 , 'PixelList');
[numofcomp,x] = size(stats);
BEcell = cell(numofcomp,2);
Ne = zeros(numofcomp,1);
Nb = zeros(numofcomp,1);
for i=1:1:numofcomp
contour = stats(i).PixelList;
S = size(contour);
b=0;
ee=0;
for j=1:1:S(1)
J = contour(j,1);
I = contour(j,2);
g1 = BW4(I,J+1) - 1;
g2 = BW4(I-1,J+1) - 1;
g3 = BW4(I-1,J) - 1;
g4 = BW4(I-1,J-1) - 1;
g5 = BW4(I,J-1) - 1;
g6 = BW4(I+1,J-1) - 1;
g7 = BW4(I+1,J) - 1;
g8 = BW4(I+1,J+1) - 1;
G1=g1*g2*g3;
G3=g3*g4*g5;
G5=g5*g6*g7;
G7=g7*g8*g1;
Nc = (g1 -G1) + (g3 -G3) + (g5 -G5) + (g7 -G7 );
if Nc < -2
b=b+1;
BEcell{i , 1}(j,1) = I;
BEcell{i , 1}(j,2) = J;
end
if Nc == -1
ee=ee+1;
BEcell{i , 2}(j,1) = I;
BEcell{i , 2}(j,2) = J;
end
end
Nb(i,1)=b;
Ne(i,1)=ee;
end
I don't know what it's wrong with it ? Have you ever seen this formula before?
also, I'm not sure about g9 there is no g9 in the matrix which is numbered by paper
EDIT
I used bwmorph :
Bifurcation = bwmorph(BW4 , 'branchpoints' ,Inf );
B=0;
for i = 1:1:Size(1)
for j = 1:1:Size(2)
if( Bifurcation(i,j)) %
B = B+1;
end
end
end
%----------------------------- Find End Points ----------------------------
Endpoint = bwmorph(can , 'endpoints' ,Inf );
E=0;
for i = 1:1:Size(1)
for j = 1:1:Size(2)
if( Endpoint(i,j))
E = E+1;
end
end
end