(Sunday, 7/12/2020) [Eastern Time (GMT-4)]
Room: AAPM ePoster Library
Purpose: This study aimed to evaluate patient positioning accuracy by resizing the scanning volume of the optical surface scanning (OSS) system and to investigate the relationship between the three-dimensional features of the surface image and the positioning accuracy for the determination of optimal scan volume with high accuracy and high time resolution.
Methods: The external contour of the head phantom was set as a reference image, and an area including the entire head was defined as a reference scan volume. The scan volume lengths in the X-axis, Y-axis, and Z-axis directions were each resized by one quarter. For each scan volume, positioning was performed using the OSS system, and the residual error was determined using cone beam CT as the gold standard. The mean absolute error (MAE) of the residual error in each scan volume was calculated to determine accuracy. We calculated the normal vector of each point in each surface image comprising a three-dimensional point cloud and investigated the relation to the positioning accuracy by histogram analysis.
Results: Reducing the scan volume of the X-axis, Y-axis, and Z-axis increased the MAE of the lateral, longitudinal, and vertical directions, respectively, and the maximum MAE for each direction was 0.58, 0.59, 0.94 mm, respectively. In such a case, the number of normal vectors with an absolute value of greater than 0.8, which indicated that there were more perpendicular surfaces to each axis, tended to be significantly less. To maintain the positioning accuracy by the OSS system, it is considered that these vectors of more than 300 points are necessary.
Conclusion: The positioning accuracy of the OSS system depends on the scan volume, and pre-evaluating normal vectors can determine the optimal scan volume.
Keywords
Optical Imaging, Image Analysis, Surface Matching
Taxonomy
IM/TH- Image Registration: General (Most aspects)
Contact Email
