Metrically independent sets in the digital plane

Abstract In computer vision a variety of metrics are used to determine the distance between lattice points. The three which are encountered most often are city block distance, chessboard distance and Euclidean distance. A set S of lattice points will be said to be (r, s)-metrically independent if the congruence of S and T under the metric r implies congruence under the metric s for every digital set T. Necessary and sufficient conditions are obtained for sets to be metrically independent with respect to the three given distances. Conditions on the interpoint distances are also determined which permit a set to be embedded in the digital plane with these metrics.

[1]  Robert A. Melter Some characterizations of city block distance , 1987, Pattern Recognit. Lett..

[2]  A. ROSENFELD,et al.  Distance functions on digital pictures , 1968, Pattern Recognit..

[3]  Ioan Tomescu,et al.  Path generated digital metrics , 1983, Pattern Recognit. Lett..

[4]  J. M. S. Simões-Pereira,et al.  On optimal embeddings of metrics in graphs , 1984, J. Comb. Theory, Ser. B.

[5]  Hans S. Witsenhausen,et al.  Minimum dimension embedding of finite metric spaces , 1986, J. Comb. Theory, Ser. A.

[6]  Frank Harary,et al.  Digital metrics: A graph-theoretical approach , 1984, Pattern Recognit. Lett..

[7]  Leonard M. Blumenthal,et al.  Theory and applications of distance geometry , 1954 .

[8]  J. Hammer,et al.  Unsolved problems concerning lattice points , 1977 .