Feasibility study on authentication based keystroke dynamic over touch-screen devices

Abstract

Currently, people uses mobile devices as a part of their lives. This leads to the risk of data leaking or stealing. Due to the security issue, there are many studies try to increase security in these devices. The authentication process was bringing to the device with hope that it will prevent users from the security's risk. Since the technology is developing fast, a new kind of risk comes to users without doubt. This study proposed a new study of feasibility to use keystroke dynamics on touchscreen devices. Using combination of four factors from keystroke dynamics, which are dwell time, interval time, interval timing ratio and distance between buttons, this study gives an accuracy at 82.18% with FAR and FRR at 0.020 and 0.178 respectively.

Improving Accuracy, Applicability and Usability of Keystroke Biometrics on Mobile Touchscreen Devices

“…In contrast to the common use of classifiers in this way [15,32,42,46], we thus strongly recommend to focus on anomaly detectors instead, when adding an implicit layer for password or PIN entry. …”

“…They reported EERs between 3.65%-7.34% with distance-based anomaly detection evaluated in a single session. Other authors used a Bayesian Net to authenticate phone users in a passcode entry task with a grid of 16 symbols [32]. They reported 82.18% accuracy with temporal features and distances between subsequent touches.…”

“…Less work has been carried out on mobile devices with touch [32,42,46,59], and without investigating touch-specific behavioural features, such as those shown in Figure 1. Recent research has successfully used similar features with gesture keyboards [11], but the potential of touch biometrics is still unknown for typing by tapping [20,55].…”

Touch gesture-based authentication: A security analysis of Pattern Unlock

“…There are three factors in authentication. 1) Knowledge factor, 2) possession factor and 3) inherence factor [9]. The knowledge factor authenticates based on something the user knows, such as PIN, Password, and Pattern.…”

“…Combining start nodes with lengths gives a microscopic view of the patterns. Table 5 shows the number of patterns that are usable when each node is chosen as the starting and having variable length (4)(5)(6)(7)(8)(9).…”