A Robust Algorithm for Voxel-to-Polygon Mesh Phantom Conversion

Justin Brown ,Wesley E Bolch,Takuya Furuta

doi:10.1007/978-3-030-21293-3_17

Justin Brown , Wesley E Bolch + Show 1 more

Open Access

https://doi.org/10.1007/978-3-030-21293-3_17

Copy DOI

Abstract

Over the past 20 years, the coupling of computational human phantoms within existing Monte Carlo radiation transport codes has required phantoms to be in a voxelized format. Very recently, however, several popular radiation transport codes such as MCNP6, GEANT4, and PHITS now facilitate direct radiation transport in phantoms that are represented by either polygon or tetrahedral mesh structures – surfaces and volumes, respectively, that define both the phantom’s body contour and its array of internal organs. While both voxel-based and mesh-based phantoms provide a high degree of anatomic realism as compared to first-generation stylized (or mathematical) phantoms, mesh-type phantoms are now considered the state of the art as they permit re-sculpting of individual organs, body regions, and overall body size and shape. They also allow for extremity articulations and inclusion of thin tissue layers of radiobiological importance. These are all features that are either not permitted in or are not readily available to voxel-based phantoms. Nevertheless, since the late 1980s, a tremendous number of voxel-based computational human phantoms have been developed from image segmentation of patient CT or MR data. As a result, there is a need for conversion of these existing voxel phantoms to mesh-type formats. The present work describes an efficient and accurate algorithm to convert voxel-based phantoms to mesh-based formats, thus permitting the user to take full advantage of these additional modeling features. For this conversion, a boundary detection algorithm was implemented in conjunction with polygon detection to form high-quality mesh data suitable for radiation transport simulation or finite element analysis. This conversion can result in a significant reduction of required simulation time and can allow current voxel data to be used in modern CAD software.

Highlights

Since their early development in the late 1950s, general-purpose Monte Carlo (MC) radiation transport codes have utilized primitive geometric structures to define material interfaces in their transport geometry, e.g., planes, spheres, ellipsoids, and truncated cones
The algorithm can convert any segmented set of voxelized data to an optimized meshed surface suitable for a variety of applications such as Monte Carlo radiation transport or finite element simulations of the interactions between electromagnetic fields and the human body, e.g., during MRI

Summary

Introduction

Since their early development in the late 1950s, general-purpose Monte Carlo (MC) radiation transport codes have utilized primitive geometric structures to define material interfaces in their transport geometry, e.g., planes, spheres, ellipsoids, and truncated cones These structures were used from the 1960s to mid-1980s to geometrically represent the human body in both its outer body contour and internal organ structure. Beginning in the late 1980s, the need for improved anatomical accuracy, along with concurrent advances in computational memory and processor speed, led to the subsequent development and use of voxel-based human computational phantoms. Those phantoms were defined by a collection of rectangular parallelepipeds (voxels) of equal or non-equal size defining each tissue material. All tissue elements within a voxel phantom are generally of uniform size and shape

Methods

Results

Conclusion