A recently developed thinning algorithm, based on computing repulsive 'forces' acting on each object pixel, produces nice skeletons, but involves some very intensive computations. As a result, the method takes a long time to thin any real image, when compared against other existing methods. It can be made practical by distributing the computation across a network of workstations. This has applications to other computationally difficult image processing and vision algorithms, and has been generalized and made relatively simple to do.