Matching sets of line segments

Abstract

We give approximation algorithms for matching two sets of line segments in constant dimension. We consider several versions of the problem: Hausdorff distance, bottleneck distance and largest common subset. We study these similarity measures under several sets of transformations: translations in arbitrary dimension, rotations about a fixed point and rigid motions in two dimensions. As opposed to previous theoretical work on this problem, we match segments individually, in other words we regard our two input sets as sets of segments rather than unions of segments.