In all the different sets of photo experiments, they noticed that the scanned volumes were much bigger compared to the control model. Initially, they thought this happened because of scaling problems. However, when they considered the thickness of the walls, they saw that the accuracy of the wall got worse nearer to the socket's base. This made them think that the main reason for the volume difference was the deterioration of the socket's inner wall. They found that the external shape was quite similar to the control model, but the inside wasn't. They used a ruler-based scaling method, which they found more accurate than the distance markers used in other research. Having several reference objects for scanning could also help if the ruler isn't visible, as they sometimes noticed in their experiments.