Automated detection and quantification of early caries lesions on images captured by intraoral camera

In recent years, quantifying the severity of early caries lesions based on white light and/or fluorescence images captured by intraoral camera has been becoming a hot spot in caries research field. In order to quantify the severity of the early caries lesions, it needs to detect and segment the caries lesions accurately first. However, to date, the published methods are mainly based on threshold techniques, and it is difficult to obtain desirable results because the intensity of the teeth changes significantly. To solve this problem, this paper presents an automated detection and quantification algorithm by using a morphological top-hat/bottom-hat method along with a multi-resolution surface reconstruction technique, which is based on local intensity and morphological characteristics of caries images captured by the intraoral oral camera, for early caries lesions. The preliminary experimental results on ex vivo data demonstrated the potential of the proposed algorithm.