ABSTRACT
ABSTRACT: Traditionally, dental anatomical reference points are chosen from a 3D digital model by the dental doctor with a mouse so that the whole process has low accuracy and repeatability. In this paper, we present a digital recognition method based on the curvature feature to automatically identify the dental anatomical reference points. First, the plaster casts are scanned to acquire the 3D digital dental models. Second, the point curvature features of the 3D dental models are obtained. Then, we use the values of Gaussian curvature kK and mean curvature kH together to identify the dental anatomical reference points. According to the result of curvature analysis, points with smaller absolute magnitude of Gaussian curvature and larger mean curvature are identified as reference points. In this way, the dental parameters are measured. To evaluate the reliability and validity of this method, a comparison between curvature analysis and manual method is presented using the intraclass correlation coecient (ICC) and the Bland-Altman statistical test. Results show that the measurement method based on the curvature feature performs well in identifying the dental anatomical reference points.
