Prediction of aesthetic reconstruction effects in edentulous patients

The aim of the study is to establish a virtual prediction method to predict aesthetic reconstruction effects in edentulous patients. The facial soft tissue surface data before and after wearing complete dentures of ten edentulous patients were acquired with a facial Three-dimension scanner. Then, the two sets of scanned data were entered into the same coordinate system. Manual interaction was performed to extract the external boundary of the perioral appearance deformation area, and the proportional relationships of key facial anatomical features were measured. A virtual prediction software module was developed based on back-propagation neural networks and a Laplacian deformation algorithm. Virtual prediction of the aesthetic reconstruction effects in the overall appearance of the lower third of the face was performed in 10 edentulous patients. The mean accuracy of the virtual predictions was approximately 0.769 ± 0.205 mm, and there were statistically significant differences between the 10 patients (p < 0.05). The scope of the changes in facial appearance of edentulous patients was smaller than the scope of the lower third of the face. This method can achieve the virtual prediction of soft tissue appearance in the lower third of the face after wearing complete dentures to an extent.