The Flexible, Extensible and Efficient Toolbox of Level Set Methods

Abstract

Level set methods are a popular and powerful class of numerical algorithms for dynamic implicit surfaces and solution of Hamilton-Jacobi PDEs. While the advanced level set schemes combine both efficiency and accuracy, their implementation complexity makes it difficult for the community to reproduce new results and make quantitative comparisons between methods. This paper describes the Toolbox of Level Set Methods, a collection of Matlab routines implementing the basic level set algorithms on fixed Cartesian grids for rectangular domains in arbitrary dimension. The Toolbox’s code and interface are designed to permit flexible combinations of different schemes and PDE forms, allow easy extension through the addition of new algorithms, and achieve efficient execution despite the fact that the code is entirely written as m-files. The current contents of the Toolbox and some coding patterns important to achieving its flexibility, extensibility and efficiency are briefly explained, as is the process of adding two new algorithms. Code for both the Toolbox and the new algorithms is available from the Web.