Publication Details
Andrey Chernikov and Nikos Chrisochoides.
Published in ACM Conference on Bioinformatics, Computational Biology and Biomedicine, pages 125 -- 134, Chicago, IL, August, 2011
Abstract
The modeling of physical processes in biomedical image analysis requires a discretization of the image space into simple shapes like triangles in two dimensions and tetrahedra in three dimensions. These discretizations are known as meshes, and the construction of the meshes as image-to-mesh conversion. There are a number of requirements on image-to-mesh conversion, the most critical of them being the shape of mesh elements in terms of the absence of small angles, the faithful geometrical representation of the tissues by the mesh elements, small number of elements for real-time Finite Element and Finite Volume analysis, and rapid execution times. We present a novel algorithm for triangular and tetrahedral image-to-mesh conversion which allows for guaranteed bounds on the smallest dihedral angle and on the distance between the boundaries of the mesh and the boundaries of the tissues. The algorithm produces a small number of mesh elements that comply with these bounds. We also describe and evaluate our implementation of the proposed algorithm on two publicly available three-dimensional medical atlases. The implementation is faster than a state-of-the art Delaunay code, and in addition solves the small dihedral angle problem.
