Geometric accuracy analysis in mesh-to-solid conversion for 3D anatomical models through Taguchi methodology

Three-dimensional reconstruction of anatomical models from clinical data faces challenges in maintaining geometric fidelityduring the conversion of surface models to parametric solids. Mesh processing techniques, such as smoothing, remeshing,and polygon reduction, aim to optimize mesh quality but can lead to loss of geometric information. A methodology wasdeveloped to process 3D surface meshes and convert them into solids from both faceted and organic geometries. Theimpact of different processing parameters on mesh topology and post-conversion geometry was evaluated using theTaguchi methodology with an L16 orthogonal design, analyzing three factors: smooth factor, remesh, and polygon reductionpercentage. The smooth factor had the most significant influence (p < 0.01), followed by remesh (p < 0.05), while polygonreduction was not statistically significant (p > 0.05). Conversion techniques produced significant differences (p = 0.0012),with faceted geometries yielding results closer to the target. Geometric deviation analysis using bidirectional Hausdorffdistance demonstrated sub-millimetric discrepancies in all cases. Faceted models showed RMS values between 0.067 and0.183 mm while organic models ranged from 0.076 to 0.225 mm with minimal deviation. Volumetric analysis revealedmoderate reductions averaging 4.53 ± 2.88% for faceted and 4.98 ± 3.07% for organic geometries without volumetricexpansion. Despite slightly higher sensitivity in organic models, deviations remained within acceptable tolerances foranatomical reconstruction and computational biomechanics, indicating preserved morphological fidelity after processing