Filtering Data from Motion Sensors with Rich Features for Monitoring Brushing Behaviors

In ambient intelligence, the potential of smart toothbrushes is often overlooked. These devices, equipped with inertial measurement units and Bluetooth Low Energy connectivity, transcend their status as mere gadgets. They can gamify toothbrushing for children or act as extensions to assistive smart homes, measuring health indicators for semi-autonomous residents. Building on this potential, our research explores a dataset from 17 participants who brushed their teeth over one week in five different locations. In this work-in-progress, We focused on applying eight additional filtering techniques to enhance the results of the original dataset. Our findings demonstrate that rigorous filtering significantly improves performance across both devices in the dataset.