Recording and Transferring Head Positions to the Virtual Head Using a Multicamera System and Laser Level

Abstract

To fulfill the requirements of computer-aided orthognathic surgery, the authors developed a method of recording head positions in pitch and roll and tested its accuracy and reliability. A laser level was used to project a horizontal laser line onto a volunteer's face. A 3-dimensional (3D) photograph of the volunteer was taken to capture the laser line using the 3dMDface System, so the head positions could be recorded. To test the accuracy and reliability of this method, 35head positions were recorded and compared with the positions recorded by the gyroscope method (Pn for pitch and Rn for roll). A cone-beam computed tomographic (CBCT) scan was taken, during which the head position was recorded by the gyroscope (P0 and R0). CBCT data were imported into ProPlan CMF 1.3 software and a virtual head was created. To reproduce each recorded head position, each 3D photograph was superimposed onto the virtual head through surface registration, and the virtual head was rotated to make the laser line parallel to the coordinate axes in the lateral and frontal views; the rotation angles were recorded, respectively, as Pn' and Rn'. Under ideal conditions, Pn' should equal Pn - P0 and Rn' should equal Rn - R0. The accuracy was evaluated using the Bland-Altman method. Reliability was tested by intraclass correlation coefficient (ICC) analysis. The 95% limits of agreement between the rotation angles recorded by the present method (Pn',Rn') and the gyroscope method (Pn - P0, Rn - R0) were -0.598° to 1.589° for pitch and -1.156° to 1.674° for roll; such a difference was generally accepted as being accurate. The ICCs were 0.996 (0.992, 0.998) for pitch and 0.998 (0.997, 0.999) for roll. The 3dMDface System and laser level method of recording head positions was accurate and reliable.