Gait-Based Smartphone User Authentication: A literature survey - 12/09/2011

Yuting Chen

Title: Gait-Based Smartphone User Authentication: A literature survey

Abstract: With the prevalence of smartphones in people’s work and personal life, the significance of smartphone security has been introduced to discussion. Among the many ways to protect smartphones, user authentication is a common and convenient approach. Unlike active methods such as PIN-based authentication that require users to perform troublesome operation constantly, biometric recognition methods provide an unobtrusive and user-friendly way to realize user authentication. Gait, the rhythmical body movements of human beings as they work, has shown by recent research to be effective as a unique personal feature. However, many problems such as the data preprocessing methods, the feature extraction theme, the construction of classifier, and the experiment setup, are still open to further discussion.

In this talk, we’ll begin by the motivation of smartphone security and some background on user authentication/classification, followed by introduction of human gait. We’ll then skim over the use of gait information as a biometric signature in academic society. Next, we’ll focus on the research that use accelerometers to collect gait information to perform user authentication on smartphones. We’ll break the authentication process into detailed stages and dig into the various techniques employed by researchers. After a brief discussion of the experimental results, we’ll conclude by providing some insights to the improvement of our team’s former research paper “Gait-based User Classification Using Phone Sensors”.

-- JiaxiJin - 09 Dec 2011