Fig 1.
The motion controller fabricated in this study.
A dry human skull was mounted on the motion controller. Four types of head motion were simulated by remote control.
Table 1.
Definition of three-dimensional landmarks used in this study.
Fig 2.
Rendered 3D surface models without motion (A) and with head motion (B-E).
Although head motion occurred during the scan, the surface models look like to the control model in appearance. A, No motion; B, right side rotation; C, left side rotation; D, upward rotation; E, downward rotation.
Table 2.
Landmark identification errors (mm) in Euclidean distance of the primary examiner for each head motion and comparison with the control model (N = 15).
Table 3.
The primary examiner’s landmark identification errors (mm) in 3-dimension of each head motion and comparison with the control model (N = 15).
Fig 3.
The primary examiner’s landmark identification errors in bilateral landmarks and comparison between the right and left sides.
A, Right side rotation; B, left side rotation; C, upward rotation; D, downward rotation. In the horizontal rotation models, the errors of left side landmarks showed higher values than the right side. In particular, Porion showed statistical significances. On the other hand, there were no significant differences between the right and left side landmarks in the vertical rotation models. *P < 0.05 by the paired t-test.
Table 4.
Interexaminer reproducibility of landmark identification errors in Euclidean distance.