Smallest Enclosing Balls of Points - Fast and Robust in C++.

(high-quality software for smallest enclosing balls of balls is available in the computational geometry algorithms library CGAL)

This is the miniball software for computing smallest enclosing balls of points in arbitrary dimensions. It consists of a small C++ header file (around 500 lines of code) and an example program that includes it and demonstrates the usage. The software comes in two versions. In the faster version, the dimension is provided as a template argument and must therefore be known at compile time: Miniball.h and miniball_example.C. In the more flexible version, the dimension is provided as a constructor argument and can therefore be determined at runtime: Miniball_dynamic_d.h and miniball_dynamic_d_example.C. The miniball software is available under the GNU General Public License. If your intended use is not compliant with this license, please buy a commercial license (&euro 300). You need a license if the software that you develop using Miniball V2.0 is not open source. Changes from previous versions. The theory behind the miniball software (Proc. 7th Annual European Symposium on Algorithms (ESA), Lecture Notes in Computer Science 1643 © Springer-Verlag, pp.325-338, 1999).

Main Features:

• Very fast in low dimensions. 1 million points in 3-space are processed within 0.5 seconds on any recent machine

• High numerical stability. Almost all input degeneracies (cospherical points, multiple points, points very close together) are routinely handled.

• Easily integrates into your code. The header is small and readable; it separates class definitions from class implementations and contains documentation for all methods of the main class Miniball.