Structure Segmentation of Dental Tissue Based on Semantic Characteristics

Abstract

Dental tissue segmentation is very challenging due to the complexity of histological structure, the variability of imaging modality and image quality. Aiming at this problem, a novel dental tissue structural segmentation method based on semantic characteristics was proposed, which not only focus on dental tissue semantic segmentation, but also the instance level tooth indication. The dental images were firstly pre-processed to decrease the impact of imbalanced local illumination. Secondly, the primary dental tissue segmentation results were obtained based on the prior knowledge of oral anatomy. Then a robust key point’s extraction strategy was investigated referring to dental tissue’s hierarchy structure. At last, the markercontrolled watershed transform was employed to complete semantic segmentation. The experimental results show that the proposed method is robust and has higher accuracy.