A method of dental model segmentation based on shortest path

Abstract

To accurately extract the gum line of teeth from the three-dimensional dental model, a new method based on the traditional Dijkstra shortest path algorithm with Sigmoid curvature transformation is proposed in this paper. The curvature of each vertex of the dental model is transformed based on the sigmoid function, and each edge of the model is assigned with corresponding weight, then the shortest path representing the gum line of teeth is calculated. Also, a queue optimization and storage structure optimization strategy is used to make the algorithm implementation efficient. The experimental results show that the proposed tooth segmentation method can extract the gum line of tooth boundary accurately.