Helical-like scan and upright CBCT imaging algorithms based on robotic-arm system

Upright position CT scans enable body imaging under realistic conditions and have been widely adopted in rehabilitation medicine. However, they face challenges such as limited acquisition angles, floor stability issues, and a restricted field of view (FoV) along the Z-axis.

The objective of this study is to propose an algorithm that enables high-quality image reconstruction in lightweight robotic-arm cone-beam CT (CBCT) systems, addressing challenges related to mechanical vibrations, a limited Z-axis scanning range, and inhomogeneous sampling.

A long Z-axis helical steel ball phantom and a greedy ball-positioning algorithm are proposed to enhance geometrical calibration accuracy. A data completeness-driven method optimizes the scanning pitch for rapid full-body scans without significant image degradation. Additionally, a normalized projection-based FDK-style algorithm enhances reconstruction quality under reverse helical scanning constraints.

The proposed upright reverse helical CBCT system demonstrated competitive reconstruction accuracy, achieving RMSE values of 0.0421 for Shepp-Logan, 0.3163 for foam-like, and 115.08 for VHP phantoms. Additionally, the proposed algorithm maintained computational efficiency, completing reconstructions in 45.3, 64.0, and 189.4 s, respectively, significantly outperforming iterative methods while preserving image quality. Furthermore, it significantly reduced radiation dose compared to conventional helical CT, achieving dose reductions from 460 to 50.7 mGy.cm for a 32 cm phantom and from 1050.0 to 112.6 mGy cm for a 16 cm water phantom.

This work provides a robust solution for long-length Z-axis imaging in upright positions, as well as for unstable and nonstandard projection sampling. The proposed framework holds potential for advancing the use of robotic-arm upright CBCT systems in orthopedic functional evaluations and other clinical applications.