A calibration procedure for center of rotation estimation in ultrasound tomography without position sensor

A calibration procedure for center of rotation estimation in ultrasound tomography without position sensor is presented. The realistic error values are reported in Monte Carlo simulation by conducting an experimental trial recording the transient movement of the calibration phantom. Simulation section is programmed to provide 180 rotations at 2° steps for a complete 360° scan. Monte Carlo simulations confirm that full 360° scan is always better than a 180° scan while some differences are revealed when comparing half-circle trajectories with different symmetry distribution. Axial position estimates are better than lateral position estimates in all cases, and again FWHM estimates uncertainties are tighter than maximum amplitude ones. Analysis of variance (ANOVA) applied to groups different from 2° group gives no statistical difference, hence all groups show no statistical difference, with the exception of the trials with only two opposite view