frechet-dist

A Clojure library to calculate the discrete Fréchet distance of two polygonal curves as stated by Eiter and Mannilla. The curves may be N-dimmensional and need not be the same length.

Usage

Simply create matrices with the ordered set of points that describe the curves. The frechet-dist function returns a hash-map with the discrete frechet distance (:dist) and the coupling sequence (:couple). The Euclidean distance is used as the default metric but you can use an arbirtrary distance function if you pass it as third argument.

(ns example.core
  (:require [frechet-dist.core :as frechet]))

  (frechet/distance [[1 2] [3 4]]
                    [[5 6] [7 8] [9 0]])
;;=> {:dist 7.211102550927978, :couple ([0 0] [1 1] [1 2])}

It is also possible to compute the partial discrete frechet distance between two curves. This is calculated as the longest section of both curves that minimizes the frechet distance. This is specially useful to extract a common section of overlaping curves. An arbirtrary distance is also optional. Please note that the partial frechet distance is not a metric function and there is no formal definition for it.

(ns example.core
  (:require [frechet-dist.core :as frechet]))

(frechet/partial-distance [[1 2] [3 4]]
                          [[1 2.5] [2.9 4.3] [5 7]])
;;=> {:dist 0.5, :couple ([0 0] [1 1])}

Additionally, it is also possible to re-sample the curves in case you might want to improve the precision of the calculation. The refine function resides in the frechet-dist.sampler namespace.

(refine [[0 3] [0 4] [0 5]] 0.3)
;[[0.0 3.0] [0.0 3.25] [0.0 3.5] [0.0 3.75] [0.0 4.0] [0.0 4.25] [0.0 4.5] [0.0 4.75] [0 5]]

License

Copyright © 2015 Camilo Roca

Distributed under the LGPL v3.