Method: A 3D structured light scanner was constructed using a pico-projector and a CMOS camera. Dental stone casts were arbitrarily mounted on an average value articulator and six scans of the anterior region were taken with the teeth together. The viewpoint for each scan varied by five degrees from the previous scan, in a horizontal arc around the labial segment. Each scan captured approximately six upper teeth and six lower teeth. Next, the models were held open on the articulator to give five millimetres of incisal separation. Six more scans were recorded from similarly varying viewpoints. From each pair of scans (teeth apart and teeth together), an axis of rotation was calculated. This was achieved by ‘virtually’ aligning the scans firstly by their mandibular teeth, and then by their maxillary teeth, to produce a transformation matrix representing the movement around the articulator hinge. A total of 36 axes were calculated by pairing the scans in all possible combinations. The locations of these 36 axes were plotted in sagittal planes located 50mm left and right of the midline, to represent the position of the temporo-mandibular joints (TMJs). The repeatability of axis location was then assessed.
Result: The average axis was calculated from the 36 individual axes. The average radius of error of the individual axes, when compared to this average axis, was 1.30mm at the level of the left and right TMJs. The standard deviation was 0.80mm. 80.6% of axes were within 2mm and 95.8% of axes were located within 3mm.
Conclusion: The repeatability of this method seems clinically acceptable. Further in vivo investigation is needed to determine whether the transverse horizontal axis on a patient can similarly be determined.