Automatic Classification and Segmentation of Teeth on 3D Dental Model Using Hierarchical Deep Learning Networks

Sukun Tian,Qing Yu,Ning Dai,Bei Zhang,Fulai Yuan,Xiaosheng Cheng

doi:10.1109/access.2019.2924262

Sukun Tian, Qing Yu + Show 4 more

Open Access

https://doi.org/10.1109/access.2019.2924262

Copy DOI

Abstract

To solve the problem of low efficiency, the complexity of the interactive operation, and the high degree of manual intervention in existing methods, we propose a novel approach based on the sparse voxel octree and 3D convolution neural networks (CNNs) for segmenting and classifying tooth types on the 3D dental models. First, the tooth classification method capitalized on the two-level hierarchical feature learning is proposed to solve the misclassification problem in highly similar tooth categories. Second, we exploit an improved three-level hierarchical segmentation method based on the deep convolution features to conduct segmentation of teeth-gingiva and inter-teeth, respectively, and the conditional random field model is used to refine the boundary of the gingival margin and the inter-teeth fusion region. The experimental results show that the classification accuracy in Level_1 network is 95.96%, the average classification accuracy in Level_2 network is 88.06%, and the accuracy of tooth segmentation is 89.81%. Compared with the existing state-of-the-art methods, the proposed method has higher accuracy and universality, and it has great application potential in the computer-assisted orthodontic treatment diagnosis.

Highlights

Segmentation of individual tooth from 3D dental models is a key technique in computer-aided orthodontic systems
We find that the network trained with the DatasetIII can get the highest values in all three measurement indexes, which indicates the accuracy of feature classification of deep learning is related to the number of training samples, and the accuracy of our network can be further improved by using the virtual sample generation technology
This paper presents an automatic segmentation and classifycation method for 3D dental model via 3D convolution neural networks (CNNs)

Summary

Introduction

Segmentation of individual tooth from 3D dental models is a key technique in computer-aided orthodontic systems. Malocclusion is a common oral disease with a high prevalence, and its prevalence is about 50% [1]. The dental models play an important role in the clinical orthodontic diagnosis. They can truly show the 3D anatomy structure of patients with malocclusion, as well as the shape and position distribution of the teeth, and assist dentists to design an efficient and accurate dental treatment plan by extracting, moving and rearranging the teeth from the dental models [2], [3]. Tooth segmentation is a core step in many oral medical research processes and is the basis for computer- aided dental diagnosis and treatment.

Methods

Results

Discussion

Conclusion