Figures index

From

Surface Registration Accuracy of Clinically Obtained Intraoral Optical Scans with Manually Threshold Segmented CBCT Data

Krzysztof Andruch, Mariusz Malecki

International Journal of Dental Sciences and Research. 2020, 8(1), 7-16 doi:10.12691/ijdsr-8-1-2
  • Figure 1. Example of the four different CBCT segmentation thresholds (Ez3DPlus software) : A – first threshold, B – second threshold, C – third threshold, D – fourth threshold.
  • Figure 2. Sample results of statistical calculations from CloudCompare collected in the form of screenshots: A – table from which the mean distance between clouds of the examined models was read, B – graph from which the Gaussian mean and standard deviation were read
  • Figure 3. The process of superimposing the model segmented from CBCT (yellow) onto the reference model obtained from the optical scan (red). A – first manual registration of the models; B – models cut off after automatic registration; C – models with contact points removed after subsequent automatic registration
  • Figure 4. The values of mean distance, standard deviation and Gaussian mean observed in the study
  • Figure 5. The correlation between the segmentation threshold and the mean distance
  • Figure 6. The correlation between the segmentation threshold and the Gaussian mean
  • Figure 7. The correlation between the segmentation threshold and the Gaussian standard deviation
  • Figure 8. An example of an object composed of a point cloud created after performing statistical calculations which shows the average distances between the examined
  • Figure 9. Visible artefacts left after segmentation and processing of models: A – STL model developed based on CBCT scan (Meshmixer); B – red model-optical scan, yellow STL model based on CBCT scan (CloudCompare)
  • Figure 10. The impact of CBCT imaging quality on the smoothness of segmented surfaces. Two images taken by different devices arranged side by side. A – Vatech Pax-I 3d Green Corea; B – PaxReve3D Corea