Tooth Condition Classification for Dental Charting Using Convolutional Neural Network and Image Processing
Arryll Mori N. Antolin, Dionis A. Padilla, Jean Marc D. Reyes
Abstract
Dental check-ups involve dental charting and are commonly performed via manual inspection only by the dentists to observe the physical properties of the teeth. The utilization of the camera enables obtaining the readily available information of the teeth visually. The proponents opted on using image processing and Convolutional Neural Network to detect the different tooth statuses used for dental charting such as sound/sealed, fillings, dental caries, indication for extraction, and missing. Existing studies utilize x-ray/ radiograph images and near-infrared imaging. The device created using Raspberry Pi Zero W and Zero pi camera as the main components acts as an intraoral camera. The tooth conditions that were identified can be viewed in the dental chart produced by the software. The proponent's dental charting system was found to be accurate after the confusion matrix for the tested data shows that it has an overall accuracy of 88.125%.