A hybrid voxel sampling method for constructing Rad-HUMAN phantom

An accurate and fast sampling method was developed on the modeling of a voxel phantom called Rad-HUMAN for radiation protection of MC-based radiation transport and simulation. The segmented organ voxels, which were assigned with three dimensional (3D) coordinates, were simplified through a two-step hybrid sampling algorithm. Firstly, certain voxels were sampled into a coordinate matrix by the nearest neighbor sampling. Secondly, the coordinate matrix was renewed using a weighted sampling. To compare visualization with the sampling, the resultant matrix was used to extract the contour of organs/tissues for constructing polygon-surface phantom. The feasibility and effectiveness of the sampling method was verified through the modeling of large organs (e.g. skeleton system) and application of transformation to MC computational geometries.