Automated segmentation of medical images for 3D printing–voxel centric processing of spatially textured voxels

Traditional medical segmentation and current 3D modelling techniques produce tissue models that capture only surface topology which lacks the volumetric material properties to accurately represent the internal structure. This paper introduces a novel voxel-based approach, Voxel Centric Processor (VCP), that leverages Hounsfield Unit (HU) values from medical imaging data to create 3D-printed models with heterogeneous exterior and interior material properties. Unlike traditional techniques that use region growth and thresholding to create homogeneous polygonal shells of the desired structure, VCP handles each voxel separately with overlapping pixel data to assign material attributes like colour, transparency, and durometer. The implementation of VCP is demonstrated for the 3D printed model of the human liver using computed tomography (CT) data through a high-fidelity tissue model that replicates the physical features of the organ, including internal structures with micron-scale resolution. Findings from this research were evaluated by employing multi-material 3D printing technologies (i.e., mutli-material jetting) to produce spatially varying material gradients and textures in 3D-printed human liver models for surgical planning. Finally, this VCP approach can process 3D digital input mapped to physical attributes that require an understanding of the internal structure for 3D-printing but could be extended to computational analysis with site-specific properties.