Distance functions defined by variable neighborhood sequences

In the field of pattern recognition, many researchers adopt the definition that distance between two points x and y on digitized space is the length of the shortest path from x to y determined by a specific sequence of neighborhood forms. The diamond distance and various octagonal distances are typical examples of these kinds of distances. However, not necessarily every sequence of neighborhood forms does define a distance function. Thus, in this paper, we present a necessary and sufficient condition for a sequence of neighborhood forms to define a distance function. Two applications of this condition are also presented.