A Computational Introduction to Digital Image Processing, 2nd Edition

Image Topology 339

(a) (b)

FIGURE 11.28: Examples of Guo-Hall skeletonization

Skeletonization Using the Distance Transform

The distance transform can be used to provide the skeleton of a region R. We apply

the distance transform, using mask 1, to the image negative, just as we did for the circle

previously. Then the skeleton consists of those pixels (i, j) for which

d(i, j) ≥ max{d(i −1, j), d(i + 1, j), d(i, j − 1), d(i, j + 1)}.

For example, suppose we take a small region consisting of a single rectangle, and find the

distance transf orm of its negative:

MATLAB/Octave

>> c = zeros(7,9); c(2:6,2:8)=1;

>> mask = [Inf 1 Inf;1 0 Inf;Inf Inf Inf];

>> cd=disttrans(1-c,mask1) ...