Fast Euclidean distance transform using a graph-search algorithm

Two new Euclidean distance transform algorithms are described. The algorithms are designed using a shortest path graph-search framework. The distance transform can be seen as the solution of a shortest path forest problem. Previous works have dealt with the Euclidean distance transform (EDT) and with the shortest path forest problem, but none of them have presented an EDT using the graph-search approach. The proposed algorithms are very simple and yet belong to the class of one of the most efficient sequential algorithms. The algorithms easily extend to higher dimensions