Tooth segmentation method based on peak clustering

Abstract

The invention discloses a tooth segmentation method based on peak clustering, and the method comprises the following steps: S1, extracting a three-dimensional dental crown model boundary, i.e., a gingival boundary B, and segmenting the B into a front boundary B2 and a rear boundary B1; S2, performing peak clustering on points on the gingival boundary B to obtain candidate tooth joint points; S3, pairing the candidate tooth joint points to obtain joint point pairs; and S4, for each joint point pair, solving a geodesic path of the joint point pair on the three-dimensional dental crown model, andsegmenting the three-dimensional dental crown model along the geodesic path between the joint point pairs to obtain a single tooth model. For the characteristics of gingival boundaries of a tooth model, density peak clustering is carried out by using the three-dimensional boundary information to obtain the joint points on the boundary as the segmentation points, the complex boundary shape similarto a sawtooth shape does not greatly influence the result by using the clustering method, the user interaction is not needed, and the tooth model can be automatically and efficiently segmented.

Description

technical field [0001] The invention relates to the technical field of graphics and image processing, in particular to a tooth segmentation method based on peak clustering. Background technique [0002] The segmentation methods for tooth mesh models can be divided into the following categories: (1) segmentation methods based on curvature (curvature); (2) segmentation methods based on projected images (image); (3) segmentation methods based on harmonic field (harmonic field) ) segmentation method. [0003] Among them, (1) the segmentation method based on curvature is as follows: [0004] The curvature-based segmentation method generally calculates the minimum principal curvature (minimum principal curvature) or mean curvature (mean curvature) of the mesh surface, and extracts the features of possible tooth boundaries (teeth-tooth boundary and tooth-gingival boundary) by setting the threshold of curvature region, since the curvature field contains a lot of useless informatio...

AI Technical Summary

Problems solved by technology

Due to the complexity of the boundary itself and using only two-dimensional information, this method may generate far more segmentation points than the actual

[0010] Yang et al. also calculated the segmentation points between adjacent teeth on the tooth model boundary, selected some control points on the boundary through user interaction, and then calculated the geodesic path between two adjacent control points to obtain a single The boundary of the teeth, good results can be obtained through user interaction, but the user needs to rotate or move the model many times to mark a specific point, too much interaction makes this method cumbersome and inefficient

Images