More flexible curve matching via the partial Fréchet similarity (Q2809313)

Measuring the similarity of curves has a large variety of applications like, among others, analysis of spectroscopic data in the context of astroinformatics and the analysis of birds' migration trajectories. In this paper a modified version of the partial Fréchet similarity is studied motivated by such applications. In such tasks of curve matching it is often necessary to ignore outliers while dissimilarities regarding individual directions should be weighted by individual costs.NEWLINENEWLINEIn this paper, beside the new extended version of the partial Fréchet similarity, a polynomial-time algorithm that computes an optimal solution of it for two given polygonal curves is also presented. The extension allows both a more flexible compartment of these curves regarding their positioning to each other and a higher potential to approximate partial Fréchet similarity under an arbitrarily chosen \(L_{p}\) metric that is in general not exactly computable over the rational numbers.