μTouch: Enabling Accurate, Lightweight Self-Touch Sensing with Passive Magnets

TL;DR

μTouch addresses the need for accurate, privacy-preserving self-touch sensing in everyday settings by leveraging passive magnets and a lightweight, self-supervised learning pipeline. The system models the measured magnetic field as $R_t = B_t + B_e + N_t$, and uses MagDelta with a 16-frame buffer to estimate the environmental field and trigger gesture windows. A three-sensor hardware layout, two passive magnet front-ends (magnetic rings and flexible magnetic silicon), and a TS2Vec-based pre-training plus three-shot fine-tuning enable rapid per-user customization. In experiments with face-touching and body-scratch tasks, μTouch achieves roughly 93–95% accuracy, remains robust after remounts and over a month, and consumes ~25 mW with ~53 ms per-sample latency, highlighting its practicality for hygiene monitoring and dermatological health.

Abstract

Self-touch gestures (e.g., nuanced facial touches and subtle finger scratches) provide rich insights into human behaviors, from hygiene practices to health monitoring. However, existing approaches fall short in detecting such micro gestures due to their diverse movement patterns. This paper presents μTouch, a novel magnetic sensing platform for self-touch gesture recognition. μTouch features (1) a compact hardware design with low-power magnetometers and magnetic silicon, (2) a lightweight semi-supervised framework requiring minimal user data, and (3) an ambient field detection module to mitigate environmental interference. We evaluated μTouch in two representative applications in user studies with 11 and 12 participants. μTouch only requires three-second fine-tuning data for each gesture, and new users need less than one minute before starting to use the system. μTouch can distinguish eight different face-touching behaviors with an average accuracy of 93.41%, and reliably detect body-scratch behaviors with an average accuracy of 94.63%. μTouch demonstrates accurate and robust sensing performance even after a month, showcasing its potential as a practical tool for hygiene monitoring and dermatological health applications.

Figures (15)

• Figure 1: Overview of $\mu$Touch. (a) Face-touching detection for infection prevention and (b) body-scratch monitoring for dermatological health. $\mu$Touch provides a unified magnetic sensing framework for accurate, privacy-preserving detection of self-touch behaviors.
• Figure 2: Pilot study of face-touching behaviors.
• Figure 3: Schematic diagram of MagDelta's detection calculations in an environmental magnetic field with and without a targeted magnet.
• Figure 4: The flowchart of the pre-train process.
• Figure 5: The flowchart of the fine-tuning process.