3D Monitoring of Toothbrushing Regions and Force Using Multimodal Sensors and Unity

Abstract

The goal of this study is to help people to monitor brushing process to maintain their oral quality by providing real-time feedback. In this research, a low-cost toothbrushing monitoring system of brushing regions and brushing force using multimodal sensors and Unity is proposed for toothbrushing quality monitoring. An inertial sensor attached to the handle of a toothbrush and Random Forester Classifier (RFC) model were used to estimate brushing regions; five force sensors clipped on the toothbrush and Random Forest Regression (RFR) model were used to estimate brushing force; a visual interface based on Unity was designed to display detection results in real-time. For brushing region detection, the results show that offline verification accuracy is 97.6%, and average accuracy of online detection method is 74.0%. For brushing force detection, 5 subjects were invited to participate in experiment on both User Dependent (UD) and User Independent (UI). The results show that average Root Mean Squared Error (RMSE) is 22.08g for UD experiment; average RMSE is 37.06g for UI experiment. For this 3D brushing monitoring system, 20 subjects were invited to participate in usability experiment. The results show that 3D brushing monitoring system of this research has good usability, performance, and user satisfaction.

Keywords

• 3D brushing monitoring system
• multimodal sensors
• machine learning
• random forest algorithm
• unity