Investigation into DCT Feature Selection for Visual Lip-Based Biometric Authentication

This paper investigated using lip movements as a behavioural biometric for person authentication. The system was trained, evaluated and tested using the XM2VTS dataset, following the Lausanne Protocol configuration II. Features were selected from the DCT coefficients of the greyscale lip image. This paper investigated the number of DCT coefficients selected, the selection process, and static and dynamic feature combinations. Using a Gaussian Mixture Model - Universal Background Model framework an Equal Error Rate of 2.20% was achieved during evaluation and on an unseen test set a False Acceptance Rate of 1.7% and False Rejection Rate of 3.0% was achieved. This compares favourably with face authentication results on the same dataset whilst not being susceptible to spoofing attacks.

Event-driven implicit authentication for mobile access control

In order to protect user privacy on mobile devices, an event-driven implicit authentication scheme is proposed in this paper. Several methods of utilizing the scheme for recognizing legitimate user behavior are investigated. The investigated methods compute an aggregate score and a threshold in real-time to determine the trust level of the current user using real data derived from user interaction with the device. The proposed scheme is designed to: operate completely in the background, require minimal training period, enable high user recognition rate for implicit authentication, and prompt detection of abnormal activity that can be used to trigger explicitly authenticated access control. In this paper, we investigate threshold computation through standard deviation and EWMA (exponentially weighted moving average) based algorithms. The result of extensive experiments on user data collected over a period of several weeks from an Android phone indicates that our proposed approach is feasible and effective for lightweight real-time implicit authentication on mobile smartphones.

Seeking help from everyone and no-one: Conceptualising the online help-seeking process among adolescent males

Online help-seeking is an emerging trend within the 21st century. Yet despite some movement towards developing online services, little is known about how young people locate, access and receive support online. This study aims to conceptualise the process of online help-seeking among adolescent males. Modified photo-elicitation techniques were employed within eight semi-structured focus group sessions with adolescent males aged 14 – 15 years (n= 56) across seven schools in Northern Ireland. Thematic analyses was conducted within an ontological framework of critical realism and an epistemological framework of contextualism. Informal online help-seeking pathways increased opportunity for social support and reduced stigma but also included loss of control and reduced anonymity. Formal pathways offered increased anonymity but concerns were raised regarding participants’ ability to locate and appraise the quality of information online. A conceptual model of online help-seeking has been developed to highlight the key help seeking pathways taken by adolescent males.