Find branches in skeleton image using matlab

Question

I have a skeleton image. I am able to detect branchpoints, endpoints and able to label image using BWMorph, BWLabel. Now I want to differntiate between skeleton center line and branches. In short I want to detect branches in skeleton. I want to perform operations on skeleton branches and I do not want to consider center line in these operations. Can someone help? Following is skeleton image. http://www.flickr.com/photos/92388309@N03/8397759970/in/photostream/ Arrows in following image show center line of skeleton. http://www.flickr.com/photos/92388309@N03/8404495186/in/photostream/

Following is code I have used.

mn=bwmorph(y,'branchpoints');
[row column] = find(mn);
branchPts    = [row column];
endImg    = bwmorph(y, 'endpoints');
[row column] = find(endImg);
endPts       = [row column];
figure;imshow(y);
hold on ; 
plot(branchPts(:,2),branchPts(:,1),'rx');
hold on; plot(endPts(:,2),endPts(:,1),'*');
% Labeling the Branches
branches = imsubtract(y,mn); % set branch points to zero
figure; imshow(branches);title('no bp')
branchesLabeled = bwlabel( branches); % label connected components
vislabels(branchesLabeled);
% Calculation of Length of Branches and Circular Neighbourhood Method for
% for detection of normal and abnormal branches.
sts = regionprops( branchesLabeled,'Area', 'Perimeter','MajorAxisLength','Centroid' );
figure
imshow(branchesLabeled)
hold on
% Loop for circles
for i=1:size(sts)
r= sts(i).MajorAxisLength/2 ; %desired radius
centerx = sts(i).Centroid(1);  
centery = sts(i).Centroid(2);
th = 0:pi/50:2*pi;
xunit = r * cos(th) + centerx;
yunit = r * sin(th) + centery;
plot(xunit, yunit);
end 

Answer 1

You may use the same technique I used in Detect Longest Vertical Lines (Edges) in an Image to locate the center line under the assumption it is the longest.

Once you find it, it is easy to locate branches.