Keystroke-Aligned Body Motion Patterns for Short-Burst Continuous Smartphone Authentication: A Proof-of-Concept Study Using Motion Capture

Continuous smartphone user authentication systems aim to verify a user’s identity by monitoring behavioral patterns during typical device usage. Current approaches based on keystroke dynamics have demonstrated robust performance over extended authentication windows ranging from 20 to 120 seconds, with recent advances reducing these requirements to 5–10 seconds. However, many common smartphone interactions, such as typing a URL in a mobile browser, entering search queries on the home screen, or composing brief text message responses, occur in very short bursts lasting only 1–2 seconds. Consequently, the current state of the art in smartphone keystroke dynamics remains unsuitable for authenticating users during these prevalent short-burst textual interactions. Building upon our previous research that demonstrated the effectiveness of integrating body motion features with swipe-based behavioral biometrics, this proof of concept study explores whether similar performance enhancements can be achieved for keystroke dynamics within short authentication windows. We investigate a multimodal approach that combines traditional keystroke dynamics with laboratory-grade 3D motion capture body movement data and smartphone motion features, utilizing keystroke events as temporal anchors for feature extraction. We evaluate our methodology on our publicly available dataset comprising of 42 users and demonstrate that the integration of phone motion and body motion features with keystroke dynamics achieves an equal error rate (EER) of 1.5% over 1-second authentication windows when utilizing all available features. This represents a substantial improvement over keystroke-only approaches, which typically achieve EERs of 9.5% to 11.9% under similar conditions. Additionally, we identify the specific body regions and feature types most valuable for short-window authentication, providing a roadmap for future implementations using smartphone sensors and wearable devices.