Image segmentation Voorbeeldclausules

Image segmentation. In order to optimize the protocol for the production of a tooth transplantation template, a comparison was made between three different segmentation methods. The first segmentation method was thresholding. Thresholding is the simplest method of image segmentation, it creates binary images from grey-level ones by turning all voxels below some threshold to zero and all voxels above that threshold to one [47]. Thresholding was performed on the 3D images of the teeth in the maxilla, but after trying on several teeth it became clear that the threshold-based method was not able to discriminate teeth from the alveolar bone. No threshold value could be found that resulted in an image of only the teeth without any alveolar bone. This can be explained by the fact that the bone regions have an intensity which is close to the intensity of teeth. In addition, the image intensity of a tooth is not homogeneously with enamel having the highest intensity and root having the lowest intensity. Figure 19 shows the result of using thresholding on randomly chosen teeth in the software program Amira. The study design (see: Figure 17) was to compare 3D images of teeth in the maxilla with 3D images of extracted teeth both segmented with the same method. Because it was not possible to segment the teeth from the maxilla using thresholding, it was of no meaning to perform thresholding on the 3D image of the extracted tooth. Therefore, the segmentation method thresholding was excluded from the study. The other two segmentation methods were both included in the software program ITK-SnAP. The region-based segmentation method tries to find coherent regions with similar pixel intensity [47]. ▇▇▇ ▇▇▇▇-based segmentation method, on the other hand, attempts to identify boundaries according to pixel differences [47]. The results of region- and edge-based segmentation on the same tooth are shown in Figure 20.