OBJECTIVE: The development of a new cranial imaging system to capture a three-dimensional (3D) model of an infant's head has been previously reported. The accuracy of this new system has been independently established. However, before replacing the traditional plaster casting technique, the two methods require a comparison to ensure that the models they produce are equivalent. METHODS: Ten sequential infants were digitized by the 3D imaging system and the plaster casting technique following previously reported protocols. The cast models were also digitized so they could be compared with the digitized images. The two models (3dImage and Cast) were then imported into dimensional analysis software and aligned and registered with well-established registration algorithms. Difference maps that identified the variation between the two surfaces were generated for each pair, and descriptive statistics of these differences were recorded. RESULTS: The mean difference between the cast and the digitized models was 0.052 mm (standard deviation = 0.988), with a root mean square (RMS) difference of 1.028 mm. Two-dimensional slices obtained from the registered 3D surfaces demonstrated excellent agreement between the cranial contours. CONCLUSIONS: In this investigation, the 3D models created by the new imaging system were found to be within 1.0 mm RMS of the models created by the plaster casting technique. Two-dimensional cranial contours demonstrated excellent agreement between the two methods. The results of this investigation confirmed that the new cranial imaging system and the traditional plaster casting technique yield equivalent models.
OBJECTIVE: The development of a new cranial imaging system to capture a three-dimensional (3D) model of an infant's head has been previously reported. The accuracy of this new system has been independently established. However, before replacing the traditional plaster casting technique, the two methods require a comparison to ensure that the models they produce are equivalent. METHODS: Ten sequential infants were digitized by the 3D imaging system and the plaster casting technique following previously reported protocols. The cast models were also digitized so they could be compared with the digitized images. The two models (3dImage and Cast) were then imported into dimensional analysis software and aligned and registered with well-established registration algorithms. Difference maps that identified the variation between the two surfaces were generated for each pair, and descriptive statistics of these differences were recorded. RESULTS: The mean difference between the cast and the digitized models was 0.052 mm (standard deviation = 0.988), with a root mean square (RMS) difference of 1.028 mm. Two-dimensional slices obtained from the registered 3D surfaces demonstrated excellent agreement between the cranial contours. CONCLUSIONS: In this investigation, the 3D models created by the new imaging system were found to be within 1.0 mm RMS of the models created by the plaster casting technique. Two-dimensional cranial contours demonstrated excellent agreement between the two methods. The results of this investigation confirmed that the new cranial imaging system and the traditional plaster casting technique yield equivalent models.
Authors: Kevin M Kelly; Edward F Joganic; Stephen P Beals; Jeff A Riggs; Mary Kay McGuire; Timothy R Littlefield Journal: Glob Pediatr Health Date: 2018-10-16