Ordered graph and ordered graph isomorphism provide a natural represen
tation of many objects in applications such as computational geometry,
computer vision and pattern recognition. In the present paper we prop
ose a coding procedure for ordered graphs that improves an earlier one
based on Eulerian circuits of graphs in terms of both simplicity and
computational efficiency. Using our coding approach, we show that the
ordered graph isomorphism problem can be optimally solved in quadratic
time, although no efficient (polynomial-bound) isomorphism algorithm
for general graphs exists today. An experimental evaluation demonstrat
es the superior performance of the new method.