A comparison of plaster, digital and reconstructed study model accuracy

Objectives: To evaluate the accuracy and reproducibility of a three-dimensional (3D) optical laser-scanning device to record the surface detail of plaster study models. To determine the accuracy of physical model replicas constructed from the 3D digital files. Design and setting: A method comparison study using 30 dental study models held in the Orthodontic Department, School of Dentistry, Cardiff University. Materials and methods: Each model was captured three-dimensionally, using a commercially available Minolta VIVID 900 non-contact 3D surface laser scanner (Konica Minolta Inc., Tokyo, Japan), a rotary stage and Easy3DScan integrating software (TowerGraphics, Lucca, Italy). Linear measurements were recorded between landmarks, directly on each of the plaster models and indirectly on the 3D digital surface models, on two separate occasions by a single examiner. Physical replicas of two digital models were also reconstructed from their scanned data files, using a rapid prototyping (RP) manufacturing process, and directly evaluated for dimensional accuracy. Results: The mean difference between measurements made directly on the plaster models and those made on the 3D digital surface models was 0.14 mm, and was not statistically significant (P=0.237). The mean difference between measurements made on both the plaster and virtual models and those on the RP models, in the z plane was highly statistically significant (P