This paper presents a pilot study on cross-device variability in online signature dynamics captured on consumer Samsung devices using S-Pen technology. Signature data were acquired on two devices, a Galaxy Ultra smartphone and a Galaxy Tab S6 Lite tablet, through a unified web-based interface designed to ensure consistent capture across platforms. The acquisition process recorded timestamped x–y trajectories, stroke events, and pressure information when available, preserving temporal structure for dynamic analysis. Genuine signatures were systematically divided into reference and test sets, and comparisons were performed under intra-device conditions (enrollment and verification on the same device) and cross-device conditions (enrollment and verification on different devices). Similarity was evaluated using Dynamic Time Warping (DTW) on multivariate time series, with analysis focused on how differences in form factor and writing area influence signature behavior. This problem is directly relevant to mobile biometric authentication workflows, where users frequently enroll on one device and later verify on another; under this mismatch scenario, reduced separability between genuine and impostor scores can affect decision reliability. Consistent with this interpretation, the results show lower dissimilarity in intra-device comparisons and higher distances with ROC degradation under cross-device mismatch. These findings provide exploratory evidence that device heterogeneity is a practical factor in mobile signature verification and support the need for cross-device-aware design in authentication systems used for digital transactions and document authorization in real-world mobile environments.
Reyes-García et al. (Thu,) studied this question.