Mask Wearing Status Estimation with Smartwatches

TL;DR

This paper introduces MaskReminder, a smartwatch-based system using deep learning to accurately detect mask-wearing status and remind users, aiding COVID-19 transmission prevention.

Contribution

It presents a novel deep learning approach with MLP-Mixer for mask-related movement recognition using smartwatch inertial data, achieving high accuracy and user-independent reminders.

Findings

89% average recognition accuracy

90% success rate in user-independent reminders

Effective detection of mask-related hand movements

Abstract

We present MaskReminder, an automatic mask-wearing status estimation system based on smartwatches, to remind users who may be exposed to the COVID-19 virus transmission scenarios, to wear a mask. MaskReminder with the powerful MLP-Mixer deep learning model can effectively learn long-short range information from the inertial measurement unit readings, and can recognize the mask-related hand movements such as wearing a mask, lowering the metal strap of the mask, removing the strap from behind one side of the ears, etc. Extensive experiments on 20 volunteers and 8000+ data samples show that the average recognition accuracy is 89%. Moreover, MaskReminder is capable to remind a user to wear with a success rate of 90% even in the user-independent setting.