Faculty Publications

Can a Smartwatch Tell if you are Brushing your Teeth Correctly?

Document Type

Conference

Keywords

machine learning, mHealth, smartwatch app, toothbrushing

Journal/Book/Conference Title

2025 IEEE International Conference on Pervasive Computing and Communications Workshops

First Page

164

Last Page

169

Abstract

Oral hygiene is a crucial part of healthcare. Unfortunately, the recent surveys suggest that less than 70% of Americans follow the guidelines of good oral health. There are so many mobile health applications but very few help in tracking toothbrushing related stats. In this work, we propose a smartwatch based app that can track how good you are brushing your teeth by leveraging only accelerometer and gyroscope sensing data. We collected toothbrushing data in real-world settings with participants wearing the smartwatch. Then we analyzed the data from the smartwatch sensors to find five crucial features that uniquely identify seven dentist-recognized toothbrushing techniques. Next, we develop machine learning models to accurately identify each of these techniques. We train five well known supervised machine learning classifiers and find that KNN works best with over 99% accuracy for six of seven toothbrushing techniques.

Department

Department of Computer Science

Original Publication Date

6-19-2025

DOI of published version

10.1109/PerComWorkshops65533.2025.00059

Link to Full Text

Comments?

Share

COinS