Verification of the tactile scanning procedure

The accuracy of the scanning procedure was verified by fitting the sampled point clouds of the femoral implant surface to the computer-aided design (CAD) files of the corresponding femoral components. Implant CAD surfaces and recorded point clouds were spatially aligned using best-fit alignment optimization of the 3D software Geomagic® Studio 2013 (3D Systems, Rock Hill, SC, USA). Reconstruction accuracy was evaluated based on the relative distance of a recorded sample point to the true implant surface. To ensure the algorithm’s function, this procedure was repeatedly performed for 21 right knees. The mean minimum and maximum deviations from the postoperatively sampled points to the respective implant surface were 0.27 mm and 1.89 mm, respectively. Concerning all patient cases, a mean distance error of 0.79 mm was determined, while the average root mean square error (RMSE) was 0.65 mm. An example of a fitted point cloud (red) to the implant surface is illustrated in Fig. 1.

Verification of the tactile scanning procedure by aligning the recorded postoperative point cloud to the CAD data of the implant surface. Exemplarily, the point cloud obtained from patient case #029 (red) aligned to a corresponding Scorpio CR femoral surface component (size 7)