Analysis on Teeth Occlusion Distribution Based on Segmentation and Registration Algorithm

Zihan Cao, Xinwu Sun, Shasha Liu, Gangyuan Chen, Yan Liu, Xinggang Liu, Dongxiang Zheng, Ling Wang

Abstract: Occlusal contact status of teeth is a key indicator for orthodontic and periodontal disease treatment. Digital analysis includes tooth position recognition and occlusal contact distribution estimation. In this study, we propose a cascade two-stage point-wise network named Teeth Segmentation Network (TSegNet) based on self-attention mechanism to address teeth segmentation task. And a template-based registration method is proposed to analyze the status of teeth occlusion. In TSegNet, spatial and channel attention are used to improve the performance feature extraction. Template-registration-based occlusal distribution analysis method reduced the labeled number of training samples. To the best of our knowledge, it is the first study on occlusal contact analyzing by using computer-aided-diagnosis technique. Experiment results illustrate the effectiveness and robustness of our proposed method.