Comparative Evaluation of Surface Determination Techniques for Dimensional Metrology under Few-View X-ray Computed Tomography

X-ray Computed Tomography (XCT) is a well-established non-destructive technique for acquiring internal and external geometrical information of components with complex structures. In practical industrial scenarios, however, the number of available projections is often limited due to constraints on acquisition time, accessibility, or radiation dose, leading to few-view reconstructions that challenge subsequent data processing. Under such conditions, surface determination becomes particularly critical, as segmentation inaccuracies are amplified and directly affect dimensional measurement reliability. This work presents a comparative evaluation of surface determination techniques (SDT) applied to XCT data reconstructed from limited projection sets. Cylindrical artifacts are employed as representative test specimens to ensure well-defined reference geometries while retaining sensitivity to reconstruction and segmentation errors. The study systematically investigates multiple acquisition scenarios, including one-quarter, one-half, and full projection configurations, enabling a quantitative analysis of segmentation performance across different levels of data incompleteness. SDTs are evaluated in terms of robustness, accuracy, and consistency under few-view conditions. The results highlight significant differences as the projection number decreases, revealing varying sensitivity to reconstruction artifacts and noise. The findings provide practical insights into the suitability of different SDTs for few-view XCT applications and contribute to improving the reliability of dimensional and geometric measurements under constrained acquisition conditions.