Handwritten Digit Leakage from Smartphone Motion Sensors Across Unseen Users and Phone Models

2026-09-26

AI Summary

This study addresses the privacy risk of handwritten digit leakage via smartphone motion sensors in unintended scenarios, where cross-user and cross-device prediction feasibility remains underexplored. Leveraging the HuMIdb dataset, this work evaluates multiple multimodal signal processing approaches, including handcrafted features paired with classical machine learning, the MiniRocket kernel, and a compact sensor Patch Transformer. The findings reveal the counterintuitive informativeness of low-motion recordings. Notably, the Transformer model achieves 57.74% and 58.77% accuracy on unseen users and unseen devices, respectively, with top-3 accuracy exceeding 82%. The implementation code has been made publicly available.

Abstract

Smartphone motion sensors support interactive applications, but their readings may also reveal touchscreen input beyond their intended use. Assuming known drawing intervals, we study whether handwritten digits remain predictable across users and devices, as a 10-class problem on 19,628 HuMIdb recordings from 481 participants. We compare handcrafted features with classical machine learning algorithms, MiniRocket kernels, and a compact sensor patch transformer on accelerometer, linear acceleration, gyroscope, and gravity signals. The transformer achieves 57.74\% accuracy and 82.64\% top-3 accuracy on 75 unseen participants, and 58.77$\pm$0.95\% over 3 seeds for unseen participants on 9 unseen phone models. Low motion recordings remain informative, accuracy is not monotonic in motion level, and the tested contrastive pretraining, augmentation, and derived signals give no consistent gains. Digits are thus predictable beyond familiar users and phone models under assumed segmentation, while acquisition-order shortcuts limit conclusions about practical privacy exposure. Code available at: https://github.com/Arritmic/motion-digit-leakage.

Problem

• motion sensors
• handwritten digit leakage
• smartphone privacy
• cross-user generalization
• cross-device generalization

Innovation

• Sensor Patch Transformer
• Side-channel Leakage
• Cross-device Generalization
• Motion Sensors
• Handwritten Digit Recognition